Академический Документы
Профессиональный Документы
Культура Документы
4.1
TECHNICAL ADDENDUM
P/N 300-007-027 REV A01
TABLE OF CONTENTS
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
Scope and Intended Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Product Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Typeface Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Notes, Tips and Warnings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
TABLE OF CONTENTS avmaint autorestart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 avmaint cancelremovebadhashes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 avmaint cat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 avmaint checkpoint. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 avmaint config . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 avmaint conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 avmaint crunch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 avmaint decommission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 avmaint ducp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 avmaint findbadhashes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 avmaint garbagecollect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 avmaint gcstatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 avmaint gendata. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 avmaint getclientmsgs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 avmaint geterrors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 avmaint getrefby. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 avmaint hfscheck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 avmaint hfscheckstatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 avmaint hfscheckthrottle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 avmaint infomessage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 avmaint kill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 avmaint logscan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 avmaint lookup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 avmaint ls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 avmaint lscp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 avmaint nodelist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 avmaint perf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 avmaint ping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 avmaint rebuildstripe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 avamaint removebadhashes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 avmaint rmcp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 avmaint sessions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 avmaint stats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 avmaint stripels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 avmaint suspend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 avmaint test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 avmaint testintegrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 avmaint timesync . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 avmaint timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 avmaint tracehash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 avmgr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 avrpm_report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 avscc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 avsetup_avamarbin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 avsetup_ems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 avsetup_mccli. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 avsetup_mcs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 avsetup_mds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 avsetup_snmp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 avsetup_webstart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 avsetup_wrapper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 avtar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 axionfs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 backup_upgrade_files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 btfix. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 AVAMAR 4.1 TECHNICAL ADDENDUM 4
TABLE OF CONTENTS capacity.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 change-passwords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 change_nodetype. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 check.dpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 check.mcs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 checklib.pm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 clean.dpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 copy-ata-drive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 copy-checkpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 cp_cron. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 cplist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 create_newconfigs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 cron_env_wrapper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 dbcreate_mds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 dbload.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 dbmaint.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 dbpurge.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 dbUpdateActivitiesExt.pl. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 delete-snapups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 dpn.pm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 dpncron.pm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 dpnctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 dpnfsctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 dpnnetutil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 dpnsummary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 dpn-time-config. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 dt and dtsh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 dump_accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 emserver.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 emwebapp.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 errchk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 evening_cron_run. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 expire-snapups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 gathergsankeydata. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 gc_cron. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 gen-ssl-cert. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 gethostbyname. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 getlogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282 getnodelogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 getsnapupstats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 health_check.pl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 hfscheck_cron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 hfscheck_kill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 hfsclean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 hfssetup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 initacnt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 java_update.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 lm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 load_accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 logmrg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 mapall. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 mcdbmaint.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302 mcdbsql.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303 mcfeature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 mcgui.bat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305 AVAMAR 4.1 TECHNICAL ADDENDUM 5
TABLE OF CONTENTS mcgui.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 mcgui_login.bat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 mcserver.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309 mcsmon_run. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 mcsnmp_cron. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 mds_ctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316 metadata_cron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317 mktime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 modify-snapups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320 monmcs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 morning_cron_run . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 nodenumbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324 opstatus.dpn. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326 pingmcs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328 probeaux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 probedump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330 probesingle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 propagate-gsan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332 psregrep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333 pull-checkpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334 rebuild.node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336 repl_cron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340 replicate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341 resite. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353 restart.dpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364 resume_crons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369 rollback.dpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370 rununtil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371 sched.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372 scn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374 showperfhistory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376 shutdown.dpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378 ssn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380 start.dpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382 start.nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391 stats.sh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395 status.dpn. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396 store-checkpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397 stunctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399 suspend_crons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401 timedist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402 timerange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405 timesyncmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406 tomcatctl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408 truncate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 ugcheck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 ugcopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410 ugprep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411 wait.crunch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417 wait.dpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419 website . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420 zzdpn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
TABLE OF CONTENTS
TABLE OF CONTENTS
FOREWORD
Scope and Intended Audience
Scope. This publication is intended to support the Avamar System Administration Manual and other publications by providing additional detailed technical information that was intentionally omitted from these publications in order to promote the best overall usability by a wide range of audiences. Intended Audience. The information in this publication is primarily intended for advanced users, EMC Field Engineers, contracted representatives and business partners.
Product Information
For current documentation, release notes, software updates, as well as information about EMC products, licensing and service, go to the EMC Powerlink web site at http://Powerlink.EMC.com.
Typeface Conventions
The following table provides examples of standard typeface styles used in this publication to convey various kinds of information.
EXAMPLE DESCRIPTION
Bold text denotes actual Graphical User Interface (GUI) buttons, commands, menus and options (any GUI element that initiates action). Also note in the second example that sequential commands are separated by a greater-than (>) character. In this example, you are being instructed to choose the Close command from the File menu.
Bold fixed-width text denotes shell commands that must be entered exactly as they appear in this publication. All caps text often denotes a placeholder (token) for an actual value that must be supplied by the user. In this example, FILE would be an actual filename. Regular (not bold) fixed-width text denotes command shell messages. It is also used to list code and file contents.
Installation Complete.
TIP: This is a tip. Tips present optional information intended to improve your productivity or otherwise enhance your experience with our product. Tips never contain information that will cause a failure if ignored.
NOTE: This is a general note. Notes contain ancillary information intended to clarify a topic or procedure. Notes never contain information that will cause a failure if ignored.
10
FOREWORD
11
Checkpoints
A checkpoint is a saved copy of the server that can be used to restart the system. A checkpoint directory is created on each active disk of each node of the system and in which are stored copies of the stripes on that disk as they were at the time the checkpoint was taken.
Checkpoints ADVANCED TECHNICAL INFORMATION Furthermore, this somewhat rigid processing model also did not take into consideration certain system states (for example, a single node offline for some reason or an error on a single disk) that might allow for usable checkpoints to be taken that would also report certain errors. For this reason, Avamar 4.0 introduced the concept of a minimal checkpoint, which is simply a checkpoint that reported one or more errors that are known to be of an acceptable nature in certain contexts. These checkpoints are not full checkpoints in the sense that they could be used to easily roll back an entire system without additional manual intervention, but neither are they completely useless. Although less desirable than full checkpoints, these minimal checkpoints might be the only checkpoints that a system is capable of producing for some period of time (for example, while a defective node is awaiting replacement or repair). Therefore, beginning with version 4.0, a new keep minimal checkpoints feature was implemented in order to control whether or not minimal checkpoints would be retained in the system. Minimal checkpoints can normally be distinguished from full checkpoints by looking at the checkpoint refcount (this measures how many nodes supply pieces of the checkpoint). If the refcount is less than the total number of nodes in the system, then the checkpoint is less than complete (not full) but might qualify as a minimal checkpoint.
Default System Behavior (Do Not Keep Minimal Checkpoints). Beginning with version 4.0, the default system behavior is not to retain (keep) anything other than full (error-free) checkpoints. Enabling the Keep Minimal Checkpoints Feature. If a decision is made to keep minimal checkpoints, you must explicitly enable the keep minimal checkpoints feature. There are two ways to do this: Persistently (for all future checkpoints until the keep minimal checkpoints feature is explicitly disabled), using the avmaint config cpkeepmin=1 command (page 76). On-demand (for a single checkpoint), by supplying the --keepmin option with the avmaint checkpoint (page 75) or cp_cron (page 224) commands. After enabling the keep minimal checkpoints feature, a single error of any kind deletes the checkpoint directories on only on the specific node that reported the error, but preserves the checkpoint directories on all other nodes, which results in a minimal checkpoint being created.
13
Checkpoints ADVANCED TECHNICAL INFORMATION Listing Checkpoints With the Keep Minimal Checkpoints Enabled Feature. Checkpoints are displayed (listed) using either the avmaint lscp or the cplist commands. These commands provide information about whether a checkpoint is valid, whether it has been validated and if it can be deleted. The mechanism for determining which checkpoints are eligible for deletion is controlled using two avmaint config parameters: cphfschecked and cpmostrecent, which control the total number of validated checkpoints and recent checkpoints that must be retained in the system at all times, respectively. Default values of cpmostrecent=2 and cphfschecked=1 specify that if a checkpoint is taken twice per day and validated once per day, only two checkpoints are retained in the system: the most recent validated checkpoint and one other. NOTE: It should be understood that checkpoint lists are not persistent. They are re-created (and checkpoints marked for deletion according to the --keepmin, cpkeepmin, cpmostrecent and cphfschecked values) each time an avmaint lscp, cplist or avmaint rmcp command is invoked.
Deleting Checkpoints With the Keep Minimal Checkpoints Enabled Feature. Automatic deletion of checkpoints occurs when cp_cron is invoked during the regular maintenance windows or explicitly using the avmaint rmcp command. In either case, a checkpoint list is constructed, checkpoints that are determined to be eligible for deletion are marked as deletable, then all checkpoints marked as being deletable are actually deleted from the system. Prior to version 4.0, a minimal checkpoint was considered equal to a full checkpoint. Version 4.0 and later systems now retain full checkpoints in favor of deleting minimal checkpoints even if they are older. This affects the selection of checkpoints that are marked to be deleted. Initially all checkpoints are marked to be deleted. Then, the list is scanned from the most recent to the oldest, and the cpmostrecent and cphfschecked full checkpoints are unmarked for deletion. If additional checkpoints can be retained, the list is rescanned for minimal checkpoints that are then handled as in previous passes through the list. The case where cpkeepmin=1 (or --keepmin=1) is similar to the default behavior except that in the first (and in this case, only) scan, all valid checkpoints, whether full or minimal, are counted and unmarked for deletion (which is the algorithm followed prior to version 4.0).
14
Operational Overview
This topic describes how to start, control (throttle), stop and get status for an HFS check. Starting an HFS Check. Administrator. HFS checks are typically scheduled by way of Avamar
You can also manually initiate an HFS check by running avmaint hfscheck (page 100) or hfscheck_cron directly from a command shell. In order for an hfscheck process to successfully start, there must be two or fewer currently running backups per storage node. For example, in a base multi-node server (with 14 storage nodes), there must be 28 or fewer (14 x 2 = 28) currently running backups in order for hfscheck process to successfully start. In a singlenode server, there must be two or fewer (1 x 2 = 2) currently running backups in order for hfscheck process to successfully start. It should also be noted that the MCS periodically flushes its database and this is in fact an avtar backup to the data server (formerly known as the GSAN). These MCS flushes are more likely to impact performance on single-node servers because the server comprises only a single node. It is allowable to start backups minutes after the hfscheck has started because the server is not in read-only mode for the hfscheck process. An HFS check once started runs to completion with no further intervention required from the operator. To determine the status of any currently-running HFS check, use avmaint hfscheckstatus. An HFS check can be stopped at any time by way of avmaint hfscheckstop (page 60). Throttling. An HFS check consumes extensive amounts of server resources. In order to reduce contention with normal server operation, an HFS check can be throttled. Refer to HFS Check Throttling (page 19) for additional information. Stopping an HFS Check. To stop a currently-running HFS check operation before it has completed, use avmaint hfscheckstop (page 60) or hfscheck_kill (page 294). Getting HFS Check Status. To return status for a currently-running or the most recently completed HFS check, use avmaint hfscheckstatus (page 105).
15
HFS check was stopped. HFS check was already running. No checkpoint ID was specified and no unchecked but valid checkpoint was found. A node (or the software within a node) died.
Re-run HFS check. Wait until completes and then rerun HFS check. Specify a checkpoint explicitly or create a new checkpoint. Re-run HFS check on the reduced nodes, or restart the node and then re-run HFS check. Determine cause, fix, and then re-run. This might be due to system overload. Reduce load on system, or restart system, and then rerun. It might also indicate serious disk corruption. Determine cause, fix, and then re-run HFS check. This error might indicate a problem with cgsan itself. Determine cause, fix, and then re-run HFS check. This might be due to system overload.
MSG_ERR_NODE_DOWN
MSG_ERR_MISC
cgsan probably died, or is otherwise not responding. Data server was unable to fork or exec the cgsan process.
MSG_ERR_FORK_FAILED
MSG_ERR_CGSAN_FAILED
MSG_ERR_TIMEOUT
Some operation such as starting cgsan did not complete within some fixed period.
16
MSG_ERR_IO_ERROR
An I/O error occurred while trying to communicate between gsan and cgsan. Preparing and starting cgsan failed for some unknown reason. Actual defects in the HFS have been detected.
This probably indicates that cgsan is in serious difficulties. Determine cause, fix, and then rerun HFS check. Verify arguments and re-run HFS check. Test error stripe integrity. Rollback to last validated checkpoint.
MSG_ERR_CMD_FAIL
MSG_ERR_HFSCHECKERRORS
Checks. a c p r
Atomic data sweeps. Composite data sweeps. Parity checks. Reference checking.
Index sweeps are run automatically if any other checks (a, c or p) are specified. A full enabling descriptor can be represented as a string. For example: h+acpr,p+acpr,u+acpr This descriptor (the default) applies all checks to all stripe classes. The following is an equivalent (and more concise) representation of the same enabling descriptor: hpu+acpr
17
Checkpoint Validation (HFS Checks) ADVANCED TECHNICAL INFORMATION Because a missing stripe class is considered to apply to all classes, and a missing check type list applies to all check types, it might also be represented by the following: hpu +acpr + Not supplying a stripe class implies that this stripe class will not be processed at all. Therefore, supplying h applies all checks to HFS stripes but none at all to persistent or accounting stripes. Note that applying a check to a stripe class does not necessarily mean that the corresponding stripes will be processed. For example, consider p+p This applies parity checking only to persistent stripes. However, because parity checking of data parity stripes requires that the corresponding data stripes have been swept (which they have not because no a or c checks have been specified), only persistent index parity stripes will be processed. Percentage of Stripes Processed. You can also constrain an HFS check to only process a percentage of the total stripes. For example: 50 50:hpu hpu:50 This setting specifies that checks be run against only 50% of all possible stripes. A single percentage only can be supplied either preceding or following the descriptor. The effect of this limitation can be understood from an analysis of how the percentage is applied: If the percentage is zero (the default) or is greater than or equal to 100, all potential stripes are processed. This is also the behavior if no percentage is supplied. Index stripes and delete stripes are always processed. If no parity checking is enabled, each node selects the desired percentage of its own data stripes and checks these. If parity checking is enabled, a subset of all eligible parity stripes system-wide is selected such that their number is simply the desired percentage of the total. Each node then extracts its own local parity stripes for processing. If data stripe checking is also enabled, data stripes are selected by looking at the parity group of each parity stripe and extracting those that are to be processed on that node. Given a large enough pool of eligible stripes, the total number of actual data stripes processed should be close to the desired percentage of the total. Note that this should ensure that all selected data parity stripes have all their data stripes swept and so parity checking can complete in the normal manner. In actual practice, requesting 50% of checked stripes does not always produce an actual count of 50% of the total; the actual figure is typically higher because it is only data stripes and data parity stripes that are affected by the specified processing percentage. However, given that the majority of the time spent in processing is in the data and parity sweeps, the total time spent (after subtracting
18
Checkpoint Validation (HFS Checks) ADVANCED TECHNICAL INFORMATION constant overheads) is close to the desired percentage of the total time to check. Therefore, a check that normally takes 12 hours to complete might be expected to take 3 hours if only 25 percent of stripes are checked. Subsequent partial checks are cumulative such that checking 50% and then an additional 50% will result in all stripes being processed. Note that if refchecks are requested, these are delayed until all stripes have been processed. If all checks on all stripes are cumulatively performed, the check itself is upgraded from partial to full. The actual percentage checked is displayed by cplist. Note that requesting 20%, then 20%, and then a further 20%, is unlikely to indicate precisely 60% of completed stripes.
DESCRIPTION
--concurrentdatastripes=NUM
Specifies maximum number of data stripes that are simultaneously processed on a single node during an index sweep. The higher the value, the more concurrent work is being performed. Default is 16 (one per disk). This option added in version 3.5. Specifies maximum number of index stripes that are simultaneously processed on a single node during an index sweep. The higher the value, the more concurrent work is being performed. Default is 1. This option added in version 3.5.
--concurrentindexstripes=NUM
19
--concurrentparitystripes=NUM
Specifies maximum number of parity stripes that are simultaneously processed on a single node during a parity sweep. The higher the value, the more concurrent work is being performed. Default is 16 (one per disk). This option added in version 3.5. IMPORTANT: Beginning with version 3.5, use of this option is discouraged in favor of --concurrentdatastripes, --concurrentindexstripes and --concurrentparitystripes. However, this option will continue to be supported in order to retain compatibility with previous releases. Specifies number (NUM) of HFS check index and parity stripes per node to check concurrently. Default is 1, which is equivalent to: --concurrentindexstripes=1 --concurrentdatastripes=16 --concurrentparitystripes=16
--concurrentstripes=NUM
--gccount
Specifies maximum number of index stripes that are simultaneously processed on a single node during the refcheck phase. This option added in version 3.5. Specifies the amount of throttled resources (measured as a percentage of CPU capacity) that can be allocated to other tasks. For example, a setting of 20 begins throttling once CPU usage exceeds 80%. Values in excess of 100 are treated as 100% (maximum throttling); a value of zero specifies that no throttling is to be performed. If a deadline is specified, the throttle level indicates a maximum allowable throttle level, and the throttle level starts at half this value. Default value is 20 if no deadline is specified; 40 if deadline is specified. This option added in version 3.5.
--throttlelevel=NUM
20
Checkpoint Validation (HFS Checks) ADVANCED TECHNICAL INFORMATION Throttle Level. Currently, two metrics are used to determine overuse of resources. The easiest to describe is the use of CPU. A number of places in the code have been CPU throttled in such a way that if over the course of some short interval (typically 200 milliseconds) CPU usage of the current thread exceeds some factor of the imposed throttle level, that thread is paused until such a time that the averaged CPU usage for that thread drops below the threshold. This has the requisite property that the more CPU used during a short interval, the longer it is paused. Note that for CPU usage, the supplied throttling level accurately provides a measure against which throttling can be applied. Therefore, a value of 50 should limit the HFS check process to about 50% of the CPU over reasonable intervals. The second, and principal, kind of throttling that is applied attempts to ensure that the total time spent processing the stripes increases linearly with the throttle level. It accomplishes this by inserting delays into each stripes processing. The result of this is that if a normal (un-throttled) HFS check takes time (T), throttling at level L will take approximately (1 + L/100)T. Throttling is controlled by the throttlelevel parameter which takes values between 0 and 100. A throttle level of 0 implies no throttling. A throttle level of 100 effectively doubles the time taken for the check. Values greater than 100 are not supported. Deadlines. A deadline can be imposed upon a check. It is provided as a number of minutes. This is not a timeout, as with garbage collection, since the HFS check does not terminate on passing the deadline. Rather it is a duration in which the checking program attempts to complete the check. It does this by varying the throttle level, the lower the level, the faster it runs, the higher the level, the slower it runs. Without a deadline argument, the throttle level is fixed by the throttlelevel parameter. However, with a deadline specified, the throttle level supplied is a maximum value. In no cases will the throttle level exceed this maximum: the maximum value therefore provides a slowest rate at which the check can take place, otherwise the check time would expand to fill the time available. Note that the server program is extremely limited on the action it can take if it estimates that the check will overrun the allotted time since it can only modify the throttle level and once this is set to its minimum value (1, not 0), no more improvement is possible. It is recommended therefore that any specified deadline be conservative, allowing for the possibility of an overrun. When no deadline is specified, the default throttle level is 20. With a deadline, the default value is 40. Single-Node Checking. This is a special feature that allows a single nodes stripes only to be checked. No checkpoint file is supplied because the check is conducted in the active stripe space. Read-only mode is imposed during this check so that backups will be suspended during the check. Note that this check is conducted by the server program, and therefore there is no delay in waiting for the checking program to start, as is normally the case. Single-node checking is enabled with --checknode=MODULE.NODE, option. For example, a single-node check of storage node 0.2 would be initiated by way of the --checknode=0.2 option. Other stripes on other nodes will also be processed if they contribute to the checked stripes. In particular, all safestripes that are part of a checked parity stripe are read so as to provide parity checking information. Currently, no parity stripes on the unchecked nodes are processed, and reference checking is disabled. AVAMAR 4.1 TECHNICAL ADDENDUM 21
Full checks are those that perform all checks on all stripes and correspond to an enabling descriptor of 100:hpu+acpr. Reduced checks are full checks that have completed on a reduced set of nodes. This occurs if a checkpoint on N nodes is checked on N-1 nodes. Single checks are those that are limited to a single node. Partial checks are all others, and occur when the stripes checked are being limited in some way either by limiting the checks being performed, or by limiting the class of stripe being checked, or by limiting the percentage of stripes being checked.
Data is corrupted only in the current version of the stripe (not the checkpoint stripe) Data is corrupted only in the checkpoint stripe (not the current version)
IMPORTANT: Reoccuring repairs, especially in the same stripe, might indicate a more serious problem. Report this behavior to EMC Technical Support. Automatic repair attempts to fix errors in all kinds of stripes (atomic data, composite data, index and parity), including headers. It uses the errors reported in the hfscheckerrors element of hfscheck.log as a guide for fixing corrupted stripes. As errors are repaired, hfscheckerrors is updated. The avmaint config option, autorepairmax=NUM, controls the behavior of automatic repair. A NUM value of 0 disables automatic repair; otherwise, NUM sets the maximum number of stripes that HFS check will attempt to automatically repair. Refer to avmaint config (page 76) for additional information.
22
Checkpoint Validation (HFS Checks) ADVANCED TECHNICAL INFORMATION hfscheck.log Results. The following is a sample segment of hfscheck.log when HFS check has either detected no errors or all errors that were detected were automatically repaired:
<hfscheckstatus nodes-queried="6" nodes-replied="6" nodes-total="6" generationtime="1186075937" checkpoint="cp.20070727130018" start-time="1186075816" end-time="1186075925" elapsed-time="109" check-starttime="1186075877" check-end-time="1186075887" status="completed" result="OK" stripeschecking="180" stripes-completed="180" type="full"> <hfscheckerrors/> </hfscheckstatus>
The following is a sample segment of hfscheck.log when stripe errors are detected but cannot be repaired:
<hfscheckstatus nodes-queried="6" nodes-replied="6" nodes-total="6" generation-time="1186061751" checkpoint="cp.20070727130018" start-time="1186061631" end-time="1186061740" elapsed-time="109" check-start-time="1186061691" check-end-time="1186061701" status="error" result="MSG_ERR_HFSCHECKERRORS" stripes-checking="180" stripes-completed="180" type="full"> <hfscheckerrors> <error stripeid="0.0-4"> <detail kind="compareindex" tag="index" hash="e9a8db5a18c60dd211381e50306f8b5e53f0a256"> <indexelem hash="0000000000000000000000000000000000000000" datastripeid="(none)" chunkid="0"/> </detail> <detail kind="compareindex" tag="index" hash="f74c4157351212413e686b63eb0fc6c213ac9920"> <indexelem hash="0000000000000000000000000000000000000000" datastripeid="(none)" chunkid="0"/> </detail> </error> <error stripeid="0.0-C"> <detail kind="comparechunk" tag="chunkdesc" hash="e9a8db5a18c60dd211381e50306f8b5e53f0a256"> <indexelem hash="0000000000000000000000000000000000000000" datastripeid="(none)" chunkid="0"/> <chunkdesc offset="4275" size="2726" type="4"/> </detail> </error> </hfscheckerrors> </hfscheckstatus>
23
Checkpoint Validation (HFS Checks) ADVANCED TECHNICAL INFORMATION The hfscheckerrors element contains one error element for each corrupted stripe. An error element contains one or more detail elements that provide additional information. Possible attributes in the detail element include the following:
ATTRIBUTE TYPE ATTRIBUTE DESCRIPTION
kind
Stripe check aborted (see log files for additional information). Data stripe compare to index stripe data failed. Index stripe data compare to data stripe failed. Data stripe or index stripe is missing a chunk. Parity stripe compare to column XOR sets failed. Index stripe element invalid or has bad owner. Stripe header invalid. Stripe change log file invalid. Stripe offline due to hard disk (media) error or missing file. Stripe ID. Chunk descriptor. Index element. Hash abbreviation.
tag
24
Checkpoint Validation (HFS Checks) ADVANCED TECHNICAL INFORMATION Other Indications. the following:
COMMAND
INDICATION
HFS check displays Result is OK (finished without errors). Checkpoint displays a triple-dash (---) (the checkpoint was not fully checked) rather than hfs. Displays number of stripes tested and repaired.
hfscheck_cron
Depending on whether automatic repair for a stripe was successful, the error log (/data01/cur/err.log) reports one of the following: hfscheck reported error, but testintegrity disagrees stripeid=STRIPE-ID or hfscheck reported error, but testintegrity cannot repair stripeid=STRIPE-ID or completed automatic stripe repair stripeid=STRIPE-ID Finally, the hfscheck log (/usr/local/avamar/var/cron/hfscheck.log) reports the following: hfscheck_cron - hfscheck tested 2 stripe(s) and repaired 2 stripe(s) Post Automatic Repair Considerations. As stated previously, automatic repair only fixes the current version of a corrupted stripe; the checkpointed version retains its original errors. So the following should be considered:
ACTION RESULTS
If a validated checkpoint is required after automatic repair, you must take a new checkpoint and run HFS check again. If the original checkpoint is later used for a system rollback, you must then run HFS check again.
Assuming all stripe errors were repaired in the previous HFS check and new errors have not been introduced, the new checkpoint should successfully validate. The system rollback will reintroduce the original errors into the current version of a stripe. The new HFS check will repair the corrupted data.
25
26
Client Agent and Plug-in Management ADVANCED TECHNICAL INFORMATION backups can still be initiated from the client provided that the client plug-in used to take that backup is not disabled. Disabling a plug-in prevents any client running that specific plug-in build from performing any backup, restore or validation activity. However, if the client agent is not disabled, the client can still communicate with the MCS. Selective Management of Plug-in Operations. System administrators can also selectively allow or disallow individual plug-in operations for all clients running a specific plug-in version (all builds) or build. These individual operations are: On-demand backups initiated from the client Scheduled backups Restores Backup validation Ability to browse stored backups on the server
27
Event Throttling
The MCS has the ability to keep track of events as they are published. If a designated number of the same event code is published in a defined period of time, the MCS will suppress displaying some of these in order to better show actual overall system activity. If this throttling did not occur, one rapidly recurring event code would overrun the Avamar Administrator event display and the system administrator would not be able to view other events. This throttling behavior is controlled by way of the following mcserver.xml settings: <root type="system"> <node name="com"> <node name="avamar"> <node name="mc"> <node name="event"> <entry key="throttle_repeat_count" value="10" /> <entry key="throttle_trigger_duration" value="10" /> <entry key="throttle_duration" value="600" /> These values are the default values for all events. throttle_repeat_count sets the number of events that must occur before throttling begins. throttle_trigger_duration sets the duration in seconds during which identical events must occur before event throttling begins. throttle_duration sets the duration in seconds for the throttling. These three values are also defined in the event catalog for each individual event. A value of DEFAULT for any of these settings in the event catalog causes the event throttler to use the default values defined in the preferences file. When an event is throttled, a new event, 22906 - EVENT_THROTTLE_START, is published indicating that event code is throttled. No more events of that particular event code are published until the throttle duration has ended. When the throttle duration ends a new event, 22907 - EVENT_THROTTLE_END is published. The event data tells the event code of the throttled event, how many events were throttled, the date/time of the first and last throttled event.
28
How OTM Works. Open Transaction Manager presents a point-in-time snapshot of a changing hard drive volume to the Avamar client by creating an alternate virtual drive. The Avamar client initializes the OTM low-level driver shortly after startup. The driver waits for a short period of inactivity (5 seconds by default) during which time no writes can occur to any of the volumes selected for backup. This is referred to as the quiescence period. Once quiescence is obtained, OTM creates a virtual drive letter for each volume selected. The Avamar backup is then created from these static virtual drives instead of the original, active volumes. OTM is a filter-class driver intercepting all read and write requests to each volume. When a write request is received, OTM pauses the write, copies the old corresponding data to the OTM cache file and then immediately writes the new data to the hard drive. This keeps the active volumes current at all times. Read requests from all applications except the Avamar client are passed directly to the hard drive without intervention. Reads from the Avamar client are trapped by the OTM filter driver which determines if the data is in the OTM cache or on the active volume. If the data is in its cache, OTM uses the cached (old) data instead of reading the data from the hard drive. OTM works at the device sector level, not at the file or filesystem level.
29
Open Transaction Manager (OTM) for Windows ADVANCED TECHNICAL INFORMATION Environment Variables and Suggested Settings. The Avamar client allows setting three OTM related values. The default settings for these parameters should be adequate in most environments.
NAME DEFAULT UNITS NOTES
freezecachesize
-100
If negative, fraction of hard drive used. If positive, hard drive space in MB.
Default of -100 indicates OTM hard drive cache will be 1/100th (1%) of the hard drive space used on the volume. (for example, cache will be 50MB on a volume with 5GB of data). The OTM cache is created on the volume being frozen. The old contents of all modified sectors are stored in the OTM cache for the duration of the backup. The OTM cache size should be increased on volumes with high activity or when backups are particularly slow. Default of 300 is 300 seconds (5 minutes). Maximum time to wait for the volume to stabilize. If the volume does not stabilize within this time, the backup might not be coherent. Higher values can cause the Avamar client to stall for extended periods if the volume never stabilizes. Default of 5 indicates there must be no writes to the volume for at least 5 seconds before the OTM snapshot can be created. Be very careful when changing this value. A value that is too low can create backups of incoherent volumes; a value that is too high might make it impossible to establish an OTM snapshot. .
freezetimeout
300
Seconds.
freezewait
Seconds.
From Avamar Administrator, these parameters can be included with on-demand backup, on-demand restore or inside a dataset using the optional plug-in command feature.
30
Retention Types ADVANCED TECHNICAL INFORMATION The optional plug-in commands feature requires that these variables be added as multiple attribute/value pairs. Each attribute must be prefixed with dash dash (--). For example:
ATTRIBUTE VALUE
-100 300 5
Retention Types
Beginning with version 4.0, all scheduled backups stored on the server are automatically and programmatically assigned one or more of the following retention types by the MCS: Daily Weekly Monthly Yearly None This is done in order to support certain advanced features, such as rendering byweek and by-month Avamar File System (AvFS) views, replicating only weekly, monthly or yearly backups, or locating certain types of backups in the Backup Management and Backup and Restore windows. When a scheduled backup job is about to start for a particular group, client or plug-in, an algorithm is invoked that calculates all applicable retention types for that backup. All applicable retention types are then included as an additional parameter in the avtar work order XML document so that the backup can be marked with the correct retention types. On-demand backups are always assigned a retention type of None. In order to discuss the algorithm used to assign retention types, consider the January and February 2006 calendar months (shown below).
31
Retention Types ADVANCED TECHNICAL INFORMATION This is a convenient example because January, 1 (that is, the first day of the first month of the year) also falls on the first day of the week (Sunday). The following table shows which retention types are programmatically assigned to various backups taken during this month: NOTE: Some backups are omitted for clarity.
RETENTION TYPE D W M Y
x x x x x x x x x x x x x x x x x x
x x x x x x x x x x x x x x x x x x X x X X
T F S S
Note that all backups are dailies; Sunday backups are also weeklies and the first daily backup of each month is a monthly. Therefore, an operation involving all monthly backups for this time period would operate on backups taken on January 1 and February 1, 2006. An operation involving weekly backups would operate on backups taken on January, 1, 8, 15, 22 and 29, 2006; as well as the backup taken on February 5.
32
33
*.cdt *.dat *.inx *.par *.pps *.rwc *.rwd *.rwi *.sps *.tab *.usd *.usi *.wlg stripeidA.stripeidB
Hash filesystem composite or directory elements. Hash filesystem data (with no hash references). Hash filesystem index. Local parity stripes. Parity of Parity stripes (parity of a far parity stripes). Persistent store composite or directory elements. Persistent store data (with no hash references). Persistent store index. Far parity stripes. Contains information about modules, nodes, tunnels and stripes. User accounting data. User accounting index. Write log file. Contains changes the have not yet been applied to a stripe. A change log file that specifies the data ranges that have been made for StripeidA by Parity StripeidB.
34
Stunnel
Stunnel allows SSL communications to be grafted on to applications that otherwise do not support secure sockets. Once configured, stunnel listens on a particular port waiting for an SSL connection. The default SSL Avamar server port is 29000, but this can easily be changed by modifying an stunnel.conf configuration file setting (page 481). Once an SSL connection has been established on the configured port, stunnel opens an outgoing unencrypted socket connection to a specified host and port, then relays data between the encrypted and unencrypted connections. Avamar implements stunnel in the following manner: Beginning with release 3.6, stunnel is the default mechanism for handling SSL communications between Avamar clients and the Avamar server, as well as among all server nodes in a multi-node system By default, stunnel is configured to listen for connections on port 29000 (the standard Avamar SSL port) and to forward connections to port 27000 (the standard unencrypted Avamar client port) The default SSL port (29000) can be changed if desired On multi-node Avamar servers, one stunnel instance will run on each storage node and on the utility node By default, the stunnel process will be automatically started when the Avamar server is started and shutdown when the Avamar server is shutdown; this behavior is controlled by way of programs that provide control of the Avamar server such as restart.dpn (page 364), shutdown.dpn (page 378), start.dpn (page 382) and start.nodes (page 391) Stunnel can also be started, stopped and configuration settings modified without affecting Avamar server operation; this behavior is controlled by way of the stunctl program (page 399)
35
36
Time Zone Application Notes ADVANCED TECHNICAL INFORMATION http://www.linux.org/docs/ldp/howto/TimePrecision-HOWTO/tz.html. This site provides information about creating and compiling your own time zone information files. zic(8). This is the man page for the time zone information file compiler, zic.
37
COMMAND REFERENCE
This chapter documents the various Avamar command-line programs, scripts and libraries.
Run Locations
Except where otherwise noted, all programs discussed in this chapter are located in the utility node /usr/local/avamar/bin directory and should be run from that location.
Wildcards
Avamar command-line utilities that accept PATTERN as a value, as well as the avtar files/directory list (page 173) can use regular expression (regex) glob operators, sometimes called wildcards, with the following limitations:
OPERATOR DESCRIPTION
Asterisk (*)
Matches zero or more occurrences of any character until the first directory delimiter character (for example, slash on Unix platforms and backslash on Windows platforms) is encountered. This effectively limits the pattern matching to a single client directory. For example, /usr/* matches the contents of /usr but not the contents of /usr/bin or /usr/local.
Matches zero or more occurrences of any character. This correlates to conventional single asterisk regex behavior. For example, /usr/** matches the entire /usr directory structure, no matter how many sub-directory levels are encountered.
Matches one occurrence of any character. Conventional regex behavior. Unlike regex, the plus sign is not processed as a glob operator; the plus sign only matches a single occurrence of the plus sign. Patterns beginning with forward slash (/) are assumed to be absolute path designations for a single directory; recursive processing of subdirectories is turned off and that directory name is not matched anywhere else. Characters enclosed in square brackets and separated by a single hyphen (-) are interpreted as a range of values. This is conventional regex behavior. For example: [0-9] matches any single numeric character [a-z] matches any single lowercase alpha character
[range of values]
In most cases, you can define multiple matching patterns in a text file and pass that file into the utility. This is generally easier than specifying multiple matches directly on the command line. However, when using a text file to pass in pattern matches, the pound sign (#) is interpreted as a comment if it appears at the beginning of a matching pattern. This will cause that entire pattern matching statement to be ignored.
Case Sensitivity. Wildcard case sensitivity will vary according to the target computing platform you are backing up. Wildcards specified for Windows platforms are not case-sensitive; wildcard specified for most other platforms are case-sensitive.
39
5minute_cron_run
The 5minute_cron_run program is primarily intended to be run as a cron job at five-minute intervals. The primary purpose is to run pingmcs (page 327) in order to verify and report on the health of the MCS.
Synopsis
5minute_cron_run [--debug] [--help]
Options
--debug --help Prints utility session information but does not actually perform the actions. Shows help, then exits.
Notes
IMPORTANT: You must load either the dpnid or admin OpenSSH key before running this utility. This program reads a default operating system user ID from the SYSPROBEUSER environment variable (page 425). If SYSPROBEUSER is not set, then remote commands are run as user admin.
40
10minute_cron_run
The 10minute_cron_run program is primarily intended to be run as a cron job at ten-minute intervals. The primary purpose is to run mcsmon_run (page 315) in order to verify and report on the health of the MCS.
Synopsis
10minute_cron_run [--debug] [--help]
Options
--debug --help Prints utility session information but does not actually perform the actions. Shows help, then exits.
Notes
IMPORTANT: You must load either the dpnid or admin OpenSSH key before running this utility. This program reads a default operating system user ID from the SYSPROBEUSER environment variable (page 425). If SYSPROBEUSER is not set, then remote commands are run as user admin.
41
ascd
The Avamar Server Communications Daemon (ASCD) facilitates communication between the storage server (also known as GSAN) and clients. Note that some services (for example, avmgr), which run on the Avamar server, are actually clients with respect to the storage server. IMPORTANT: ascd is not intended to be run directly from the command line. It is typically invoked by other programs such as restart.dpn (page 364) or start.dpn (page 382).
Synopsis
ascd [--autorestart] [--basedir=PATH] [--cleanup] [--forcestatus] [--hfsport=PORT] [--keyfile=PATH] [--norun] [--scanfile=PATH] [--sslport=PORT] [--verbose]
Options
--autorestart Turns on (enables) the automatic node restart feature. Refer to avmaint autorestart (page 71) for additional information about the automatic node restart feature. This option added in version 4.1. --basedir=PATH --cleanup --forcestatus Specifies full PATH of ascd top-level (root) directory. Default is /usr/local/avamar. If supplied, a new log file is started. This is the default setting. Forces perceived status of all storage server (also known as GSAN) nodes to be up (fully operational). This is not the default. This option added in version 4.1. --hfsport=PORT --keyfile=PATH Sets Avamar server PORT number. Default value is 27000. Specifies full PATH and filename of license file. Default is /usr/local/avamar/etc/license.xml. Refer to license LICENSEFILE (page 60) for additional information. Does not actually execute commands. Specifies full PATH and filename of the enhanced server log scanning parameters information file, which is created using the avmaint logscan command. Refer to avmaint logscan (page 113) for additional information. This option added in version 3.7.1. --sslport=PORT Specifies SSL data PORT. Default is 29000. Refer to Stunnel (page 35) for additional information. --verbose Provides maximum information. AVAMAR 4.1 TECHNICAL ADDENDUM 42
--norun --scanfile=PATH
asktime
The asktime program interactively configures Network Time Protocol (NTP) for an Avamar server.
Synopsis
asktime [--debug] [--help] [--probedir=PATH] [--verbose]
Options
--debug --help --probedir=PATH --verbose Prints utility session information but does not actually perform the actions. Shows help, then exits. Specifies parent directory for probe.out. Provides maximum information.
Notes
IMPORTANT: This utility must be run as user dpn; This utility attempts to load the dpnid OpenSSH key from the dpn user .ssh directory. This utility will fail if it is run as a different user. If asktime is run as user admin (despite being warned to quit and to run asktime as user dpn instead), then permissions problems might occur that could prevent time configuration files from being properly distributed. This only happens if asktime was previously run as user dpn and then asktime is later run as user admin. The solution is to apply chmod -R ug+rw /usr/local/avamar/var/time-config-files before running asktime as user admin. IMPORTANT: Before making any changes to your server NTP settings, shut your server down using the dpnctl stop command. On multi-node servers, asktime must be run from the utility node. This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
43
Files Produced
mktime.custom - customized script to produce localized ntpd config files asktime.answers - file containing user responses asktime.log - log file asktime offers external servers automatically only if DNS IN A resource records exist for the name ntp in the DNS zone of the utility node. The following log files are of interest on the utility node, or on single-node Avamar servers: /usr/local/avamar/var/asktime.log /usr/local/avamar/var/timedist.log /usr/local/avamar/var/timesyncmon.log /usr/local/avamar/var/cron/ntpd_keepalive_cron.log The following log files are of interest on the storage nodes in multi-node Avamar servers; /usr/local/avamar/var/timesyncmon.log /usr/local/avamar/var/ntpd_keepalive_cron.log This application automatically rotates its log file whenever it is run and when the log file exceeds 1MB in size. Up to eight log files are retained. In a multi-node Avamar system with no external time servers, two nodes are used as time servers internal to the Avamar system: 0.s and 0.0. Isolated from external time servers, node 0.s uses its local system clock, and all other nodes depend on that isolated system time. This isolated system time might therefore drift from real time, but all nodes should nevertheless synchronize to the time on the utility node. External time servers must offer NTP (network time protocol) as described in RFC 1305. This TCP/IP network service uses port 123/udp.
Processing Flow
1. Ask about external time servers. (a) Can asktime programmatically find any previously specified external time servers? No: Go to step 1(b). Yes: Use these previously specified external time servers? Yes: Go to step 2. No: Go to step 1(b).
44
asktime COMMAND REFERENCE (b) Can asktime programmatically determine if 'ntp' resolves in DNS? No: Go to step 1(c). Yes: Use these NTP servers? Yes: Go to step 2. No: Go to step 1(c). (c)Do you wish to use U.S. public Internet time servers? No: Go to step 1(d). Yes: Please enter the name of a U.S. time zone Set $public_server for use in step 3(b). Go to step 1(d). (d)Are there other external time servers that you would like to use? No: Go to step 2. Yes: Please enter the IP addresses of the external time servers Go to step 2. 2. Ask about module addresses. asktime must determine some important network configuration information before proceeding any further. In a multi-node system, asktime grants use of the internal time servers (0.s and 0.0) only to other Avamar nodes. In a dedicated subnet, the entire subnet can be configured to have access to the internal time servers. In a non-dedicated subnet, permission must be granted to each node individually. In the latter case, ntpd configuration must be updated each time a node is added (a) Can asktime programmatically determine if this a single-node server? Yes: Go to step 3. No: Do you have a dedicated Avamar subnet configuration? No: Go to step 3. Yes: Go to step 2(b). (b) Can asktime programmatically find any previously specified subnets? No: Go to step 2(c). Yes: Use these previously specified module subnets? No: Go to step 2(c). Yes: Go to step 3. (c)This host, <host-name>, is a member of subnet X.Y.Z.N/W. Use this subnet? Yes: Go to step 3. No: Please enter the subnet addresses of the other modules. Go to step 3.
45
asktime COMMAND REFERENCE 3. Ask about local time zone. (a) Can asktime programmatically determine a previously specified local time zone? No: Go to step 3(b). Yes: Use this previously specified local time zone? Yes: Go to step 4. No: Go to step 3(b). (b) Is $public_server set from step 1(c)? Yes: Go to step 4. No: Go to step 3(c). (c)You will need to choose a local time zone... browse? No: Go to step 3(d). Yes: Was a valid local time zone selected? No: Go to step 3(d). Yes: Go to step 4. (d)Please enter the name of the local time zone. Go to step 4. 4. Ask about nodes to configure. (a)Do you wish to configure the default set of nodes? Yes: Go to step 5. No: Please enter the list of nodes to be configured Go to step 5. 5. Ask whether to proceed with installation of configuration files. (a)Do you wish to proceed with installation? Yes: Go to step 6. No: Procedure complete; asktime exits. 6. Ask about current time. (a)We need an approximately current initial date and time... Is this approximately correct? Yes: Go to step 7. No: Enter current time. Go to step 7. 7. Ask whether to wait and watch for time synchronization. (a)Do you wish to wait and watch for time synchronization? Yes: Go to step 7(b). No: Procedure complete; asktime exits.
46
asktime COMMAND REFERENCE (b)We appear to have time synchronization... Do you wish to see the results? No: Procedure complete; asktime exits. Yes: Show time synchronization results. Procedure complete; asktime exits.
Troubleshooting Information
The following table describes how to recover from common problems encountered when running asktime.
ERROR/SYMPTOM DESCRIPTION/REMEDY
Verify that you are using the proper Avamar probe utility by entering: which probe /usr/local/avamar/var/probe should be returned. Verify that $SYSPROBEDIR is either not set or is set to the default probe.out location ( /usr/local/avamar/var/) by entering: echo $SYSPROBEDIR Re-run probe and asktime by entering: probe AVAMARSERVER asktime Where AVAMARSERVER is the Avamar server hostname as defined in corporate DNS.
Any error involving a missing program. Your response, ..., was invalid
Ensure that your PATH variable includes /usr/local/avamar/bin/. This error is in response to a user entry that appears to be incorrect. Common reasons for these errors are: If entering a subnet, ensure that you enter a full dotted quad (for example, 10.0.50.0) and not just the first few octets. If entering a subnet, ensure that you enter a width specifier (for example, /24, as in 10.0.50.0/24) and that the width specified is reasonable and correct. Reasonable values range from /24 to /29; anything outside that range is suspicious, although asktime will accept anything in the range /1 to /31. If entering a time zone specifier, ensure that you enter the case-sensitive name of a file that actually exists in /usr/share/zoneinfo/. To determine which files are available, enter ls -CRF /usr/share/zoneinfo in another command shell.
47
48
avacl
The avacl utility locates the .avamarmetafile file and restores encapsulated filesystem metadata to restored files and directories. The metadata content varies by filesystem. It can include Access Control Lists (ACLs), alternate data streams, and system-defined or user-defined extended attributes. Avamar File System (AvFS) contains avamarbin directories at the root directory of each client name. Each avamarbin directory contain platform-specific subdirectories (for example, aix5, hpux11, macosx, rh7.1, rh9, rhel3, rhel3-64, sol6, sol8, win32 and so forth) that contain platform-specific avacl binaries. You should save the client avamarbin directory to tape any time you archive an Avamar backup. Similarly, you must use the correct avacl binary for your specific platform when restoring filesystem metadata. The avacl utility can only restore like ACLs. For example, if you archived Sun Solaris backups to tape, you can only restore Sun Solaris ACLs. avacl does not have any ability to translate those ACLs to equivalent ACLs on another computing platform.
Synopsis
avacl [--delete] [--noacls] [--norecurse] [--quiet] [--rootdir=PATH] [--version]
Options
--delete --noacls --norecurse --quiet --rootdir=PATH --version Deletes .avamarmetafile files after applying all of the ACLs. Suppresses the application of ACLs. This is not the default setting. Does not recursively process subdirectories (processing is limited to the root directory). This is not the default setting. Runs quietly (suppresses status messages). Specifies root directory to start processing. Default is current working directory. Shows version, then exits.
Notes
This program added in version 3.7. Beginning with version 4.1, filesystem metadata is stored in a single .avamarmetafile file instead of many .avamarmetadata subdirectories. For backward compatibility, avacl is still able to process .avamarmetadata subdirectories if they are found. If both .avamarmetafile files and .avamarmetadata subdirectories are present, then .avamarmetafile will be used. The avacl utility can only run on clients running version 4.1 or later Avamar software.
49
avagent.bin
The avagent.bin program is the client background process that passes backup and restore requests to the Avamar server. IMPORTANT: On Microsoft Windows platforms, avagent.bin must run as a non-interactive service. Therefore, command line interaction with avagent.bin is not supported on Microsoft Windows platforms.
Synopsis
avagent.bin [--daemon] [--dpnacl=USER@AUTH] [--dpndomain=PATH] [--flagfile=FILE] [--help] [--informationals] [--initialize] [--listenport=PORT] [--logfile=FILE] [--mcsaddr=IP-ADDR] [--mcsport=PORT] [--nostdout] [--nowarnings] [--quiet] [--register] [--service] [--uninitialize] [--usage] [{--verbose | -v}] [--version ]
Options
--daemon --dpnacl=USER@AUTH Runs this utility as a Unix daemon. Unix platforms only. Default value is TRUE. Specifies the Avamar user ID (account name) during client registration. Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --dpndomain=DOMAIN --flagfile=FILE Specifies client home DOMAIN in the Avamar server during client registration. Specifies a FILE containing a list of options and values that will be processed by avagent.bin as if they were included on the command line. Default is /usr/local/avamar/etc/usersettings.cfg. Shows full online help (commands, options and full descriptions) for this utility, then exits. Turns on all status messages. Registers this client with the Avamar server. This only needs to be performed once. Subsequent calls are ignored. IMPORTANT: This command requires Avamar superuser privileges (page 151). Same as --register.
50
avagent.bin COMMAND REFERENCE --listenport=PORT --logfile=FILE --mcsaddr=IP-ADDR --mcsport=PORT Specifies data PORT that avagent.bin will use to listen for MCS pages. Enables information logging to this FILE. If no FILE value is supplied, default log file (avagent.log) is used. Specifies MCS IP address (IP-ADDR). Specifies data PORT that avagent.bin will use to communicate with the MCS. Default setting is port 28001. Turns off output to stdout. However, if --logfile=FILE is supplied, output will still go to log file. --nowarnings --quiet --register Turns off warning messages. Turns off both warnings and status messages. Equivalent to --noinformationals plus --nowarnings. Registers this client with the Avamar server. This only needs to be performed once. Subsequent calls are ignored. IMPORTANT: This command requires Avamar superuser privileges (page 151). Same as --initialize. --service --uninitialize --usage --verbose | -v --version Runs this utility as a Windows service. Windows platforms only. Default value is TRUE. Un-registers this client with the Avamar server. Shows abbreviated online help (commands and options only, no descriptions) for this utility, then exits. Supplying either --verbose or -v turns on all messages (status and warnings). Shows version, then exits.
--nostdout
51
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
--browse-nfs
Enables browsing of network mounted filesystems. By default, avagent.bin does not traverse network mounted filesystems when browsing for files and directories. This option added in version 2.0.
--threadstacksize=SIZE
Explicitly sets the default stack size. Default value is zero (0), which specifies that the operating system default stack size should be used. This option added in version 3.5.
52
avmaint
The avmaint program is a command-line maintenance and statistics gathering utility for the Avamar server. The avmaint program is highly complex, with many powerful commands. Therefore, for the sake of clarity, many avmaint commands are documented as if they were entirely separate utilities. Furthermore, only options that are truly global in nature (that can be supplied with any avmaint command) are documented in this topic. Options that are specific to one or more commands are documented with those commands.
Synopsis
avmaint {autorestart | cat ATOMIC | convstatus | cpstatus | datacenterlist | ducp CP-ID | gcstatus | getclientmsgs | geterrors NODE-ID | hfscheckstatus | kill SESSION-ID | ls LOCATION | lscp | nodelist | perf {reset | status} | ping | resume | rm PATH | sessions | stat OBJECT | stats | stripels STRIPELIST | suspend | tunnellist} [{--account=LOCATION | --acnt=LOCATION}] [--ap=PASSWORD] [--flushDB] [--forceaddr] [--help] [--helpxml] [--hfsaddr=AVAMARSERVER] [--hfsport=PORT] [--id=USER@AUTH] [--informationals=N | --noinformationals] [--log] [--logfile=FILE] [--maxlines=NUM] [--path=LOCATION] [--pswd=PASSWORD] [--quiet] [--tryagain] [--usage] [{--verbose | -v}] [--version] [--xmlperline=NUM]
Commands
One (and only one) of the following commands must be supplied on each command line: autorestart Controls the Avamar automatic node restart feature. Refer to avmaint autorestart (page 71) for a complete discussion of avmaint autorestart behavior and output. This command added in version 4.1. cat ATOMIC Lists contents of persistent ATOMIC. Refer to avmaint cat (page 74) for a complete discussion of avmaint cat behavior and output. convstatus Returns cumulative status of server stripe conversion operations. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. This command added in version 3.5.
53
avmaint COMMAND REFERENCE cpstatus Returns status of currently running or most recently run checkpoint. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. This command added in version 3.0.1. datacenterlist Lists Avamar modules (datacenters). Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. ducp CP-ID Lists hard drive space consumed by a specific checkpoint (CP-ID). Refer to avmaint ducp (page 90) for a complete discussion of avmaint ducp behavior and syntax. gcstatus Returns status for an on-going or the most recently completed garbage collect process. Refer to avmaint gcstatus (page 95) for a complete discussion of avmaint gcstatus behavior and syntax. getclientmsgs Returns contents of the client message store. Refer to avmaint getclientmsgs (page 97) for a complete discussion of avmaint getclientmsgs behavior and syntax. geterrors NODE-ID Return errors from the specified NODE-ID, which can be obtained by using the nodelist command (page 122). Refer to avmaint geterrors (page 98) for a complete discussion of avmaint geterrors behavior and syntax. hfscheckstatus Returns status of a currently running HFS check operation. Refer to avmaint hfscheckstatus (page 105) for a complete discussion of avmaint hfscheckstatus behavior and syntax. Also refer to Checkpoint Validation (HFS Checks) (page 15) for additional detailed technical information about the HFS check feature. kill SESSION-ID Stops (kills) this client SESSION-ID. Refer to avmaint kill (page 112) for a complete discussion of avmaint kill behavior and syntax. ls LOCATION Lists all objects in this LOCATION and below. Refer to avmaint ls (page 118) for a complete discussion of avmaint ls behavior and syntax. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. This command added in version 2.0.2.
54
avmaint COMMAND REFERENCE lscp Lists available checkpoints. Refer to avmaint lscp (page 119) for a complete discussion of avmaint lscp behavior and syntax. nodelist Returns status of all nodes. Refer to avmaint nodelist (page 122) for a complete discussion of avmaint nodelist behavior and syntax. perf {reset | status} avmaint perf status returns operational status for the entire server or specified nodes. avmaint perf reset resets the performance monitoring statistics. Refer to avmaint perf (page 124) for a complete discussion of avmaint perf behavior and syntax. ping Returns stripe status. Refer to avmaint ping (page 128) for a complete discussion of avmaint ping behavior and syntax. resume rm PATH Restarts (resumes) suspended Avamar client sessions. Removes specified persistent store map element, where PATH is the location of a persistent store map element on the server. This command added in version 3.0.1. sessions Returns information on the current active client sessions; output displays in XML format. Refer to avmaint sessions (page 133) for a complete discussion of avmaint sessions behavior and syntax. stat OBJECT stats Returns status of persistent OBJECT. Returns status of all stripes by sending a message to each stripe and receiving back some information about each stripe that responds. Refer to avmaint stats (page 134) for a complete discussion of avmaint stats behavior and syntax. stripels STRIPELIST Displays file data for specified list of stripes (STRIPELIST). Refer to avmaint stripels (page 136) for a complete discussion of avmaint stripels behavior and syntax. suspend Temporarily halts client activities and denies new client requests. Refer to avmaint suspend (page 137) for a complete discussion of avmaint suspend behavior and syntax.
55
avmaint COMMAND REFERENCE tunnellist Lists the tunnels. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML.
Global Options
--account=LOCATION | --acnt=LOCATION Supplying either --account=LOCATION or --acnt=LOCATION specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. Same as --path=LOCATION. --ap=PASSWORD PASSWORD for the --id=USER@AUTH account. Same as --pswd=PASSWORD. --flushDB --forceaddr --help --helpxml Flushes debug output after every message. Forces use of values specified by the --hfsaddr and --hfsport options. Shows full online help (commands, options and full descriptions) for this utility, then exits. Shows full online help (commands, options and full descriptions) for this utility in XML format , then exits. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). This value is typically recorded in a configuration file. It can also be set by way of an environment variable or on the command line with each transaction. Default address is localhost (127.0.0.1). Same as --server=AVAMARSERVER. --hfsport=PORT --id=USER@AUTH Sets Avamar server PORT number. Default value is 27000. Authenticate as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar.
--hfsaddr=AVAMARSERVER
56
avmaint COMMAND REFERENCE --informationals=N | --noinformationals Supplying --informationals=N sets information level. For example, --informationals=3. Supplying --noinformationals turns off all status messages. --log --logfile=FILE --maxlines=NUM Redirects and appends output to an alternative log file specified by the --logfile=FILE option. Used with the --log option to specify the full path and filename of the alternative log file. Specifies maximum number (NUM) of lines to write to the log file or number of log entries to return. Default value is zero (i.e, return all lines or entries). --path=LOCATION Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. Same as --acnt=LOCATION. --pswd=PASSWORD Specifies PASSWORD for the --id=USER@AUTH account. Same as --ap=PASSWORD. --quiet --tryagain --usage --verbose | -v --version --xmlperline=NUM Suppresses all debugging messages. If server is idle, try completing this operation later. This is the default setting. Displays a list of available options, then exits. Supplying either --verbose or -v turns on all messages (status and warnings). Shows version, then exits. Controls number (NUM) of XML attributes per line.
57
Avamar-Only Commands
Avamar-only commands access advanced functionality that is normally reserved for use by expert users only. Some of these advanced commands require that you include the --avamaronly option. IMPORTANT: Misuse of these advanced commands can cause loss of data. If you are unsure about any aspect of these commands, contact EMC Technical Support for additional information before using them.
access MODE
Sets server access MODE to one of the following: admin normal readmigrate readonly sync
Refer to avmaint access (page 69) for a complete discussion of avmaint access behavior and syntax. cancelremovebadhashes Cancels a pending find/delete bad hash operation and returns the server to a fully operation state. Refer to avmaint cancelremovebadhashes (page 73) for a complete discussion of avmaint cancelremovebadhashes behavior and output. This command added in version 4.1. checkpoint Starts a checkpoint process. Refer to avmaint checkpoint (page 75) for a complete discussion of avmaint checkpoint behavior and syntax. config Configures various internal characteristics of the server. If avmaint config is run without parameters, it returns a current list of server configuration settings. Refer to avmaint config (page 76) for a complete discussion of avmaint config behavior and syntax. conversion Converts older format data stripes to latest version. Refer to avmaint conversion (page 86) for a complete discussion of avmaint conversion behavior and syntax.
58
avmaint COMMAND REFERENCE crunch ACTION Performs one of the following ACTIONs: limit status reset rollover resetandrollover factoryreset
Refer to avmaint crunch (page 87) for a complete discussion of avmaint crunch behavior and syntax. decommission Decommissions a storage node or an individual disk within a storage node. Refer to avmaint decommission (page 89) for a complete discussion of avmaint decommission behavior and syntax. findbadhashes Returns a list of hashes and backups that are suspected to be bad, as well as a verification token, which must be supplied to removebadhashes (page 131) in order to actually remove bad hashes from the server. Refer to avmaint findbadhashes (page 91) for a complete discussion of avmaint findbadhashes behavior and syntax. This command added in version 4.1. garbagecollect Starts a garbage collect process. Refer to avmaint garbagecollect (page 93) for a complete discussion of avmaint garbagecollect behavior and syntax. gckill gendata Stops (kills) a garbage collect process. Generates data chunks for testing purposes. Refer to avmaint gendata (page 96) for a complete discussion of avmaint gendata behavior and syntax. getrefby TYPE HASH Returns all hashes of a particular TYPE that reference this HASH. Refer to avmaint getrefby (page 99) for a complete discussion of avmaint getrefby behavior and syntax. hfscheck Initiates an HFS check operation. If avmaint hfscheck is run without options, an HFS check is performed on the most recent unvalidated checkpoint file. Refer to avmaint hfscheck (page 100) for a complete discussion of avmaint hfscheck behavior and syntax. Also refer to Checkpoint Validation (HFS Checks) (page 15) for additional detailed technical information about the HFS check feature.
59
avmaint COMMAND REFERENCE hfscheckstop Terminates (stops) a currently-running HFS check operation. Refer to Checkpoint Validation (HFS Checks) (page 15) for additional detailed technical information about the HFS check feature. Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. hfscheckthrottle Overrides throttle values for an in-process HFS check operation. Refer to avmaint hfscheckthrottle (page 109) for a complete discussion of avmaint hfscheckthrottle behavior and syntax. Also refer to Checkpoint Validation (HFS Checks) (page 15) for additional detailed technical information about the HFS check feature. infomessage TEXT Writes this TEXT message to the server log. Refer to avmaint infomessage (page 111) for a complete discussion of avmaint infomessage behavior and syntax. init {admin | fullaccess} Sets server run level to one of the following: admin Administration-only mode (only root and aroot users can access the server). Full access by all users.
fullaccess
IMPORTANT: Beginning with version 2.0.2, support for numeric run levels 4 and 6 has been discontinued in favor of using admin and fullaccess string names, respectively. license LICENSEFILE Creates new Avamar server license from LICENSEFILE, where LICENSEFILE is the full path to a valid Avamar server license file. /usr/local/avamar/etc/license.xml is the default license filename and location. You must include the --avamaronly option in order to use this command. This command added in version 3.5.
60
avmaint COMMAND REFERENCE logscan [FILE] Enables ehanced server server log scanning, which will detect and report certain kinds of hardware failures. Refer to avmaint logscan (page 113) for a complete discussion of avmaint logscan behavior and syntax. This command added in version 3.7.1. lookup [H | P | U] HASH Looks up the supplied HASH and returns information about the associated index and data stripes. Refer to avmaint lookup (page 117) for a complete discussion of avmaint lookup behavior and syntax. nodestate NODE-ID=STATE Sets this NODE-ID to this STATE. This command added in version 2.0.2. rebuildstripe {STRIPE-ID | --full} Rebuilds offline stripes. This command accepts either a STRIPE-ID or the --full option. Refer to avmaint rebuildstripe (page 130) for a complete discussion of avmaint rebuildstripe behavior and syntax. removebadhashes Removes a set of hashes that are tied to a specific verification token. Refer to avamaint removebadhashes (page 131) for a complete discussion of avmaint removebadhashes behavior and syntax. This command added in version 4.1. rmcp {CP-ID | --full} Removes one or more checkpoints. Refer to avmaint rmcp (page 132) for a complete discussion of avmaint rmcp behavior and syntax. rmhash [H]HASH ... Removes one or more unreferenced hashes, where HASH is a 20-byte SHA-1 hash. Multiple hashes are separated by white space. You must include the --avamaronly option in order to use this command. This command added in version 3.6.
61
avmaint COMMAND REFERENCE shutdown {--force | --now} Shuts down Avamar server. Two modes are available: --force Causes all processes to stop immediately without sending any special information to the current client sessions. This will interrupt current users of the system. Gracefully stops the Avamar server. Connected clients receive a SERVER_EXITING error immediately prior to the server closing the connection socket. This is the default mode if no options are supplied.
--now
stripestate STRIPE-ID=STATE
Sets this STRIPE-ID to this STATE. This command added in version 2.0.2.
test
Simulates various internal server hardware faults for testing purposes. Refer to avmaint test (page 138) for a complete discussion of avmaint test behavior and syntax.
testintegrity STRIPE-ID
Tests integrity of STRIPE-ID and its parity group. Status for each of the stripes in the parity group is returned; returned status codes will be good, corrupt or unknown. Refer to avmaint testintegrity (page 139) for a complete discussion of avmaint testintegrity behavior and syntax.
timesync
Synchronizes server times with client. Refer to avmaint timesync (page 140) for a complete discussion of avmaint timesync behavior and syntax.
timing OPERATION
Refer to avmaint timing (page 141) for a complete discussion of avmaint timing behavior and syntax. tracehash HASH Traces the specified HASH. Refer to avmaint tracehash (page 143) for a complete discussion of avmaint tracehash behavior and syntax.
62
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by expert users only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
--acport=PORT
Specifies client agent data port number. Default is 28002. Same as --listenport=PORT.
Enables Avamar-only commands (page 58) and options (page 63). Use active connection load balancing. Sets local cache directory. Default value is /usr/local/avamar/var. Same as --vardir=DIR.
Specifies data port plug-ins will use to communicate with parent. Default is 28002. If not empty, use this control interface NAME specified in plug-ins. Specifies character encodings. Sets encryption method. Valid settings are: proprietary ssl:AES128-SHA ssl:AES256-SHA ssl No encryption. 128-bit Advanced Encryption Standard. 256-bit Advanced Encryption Standard. A special mode in which avmaint negotiates and uses the strongest encryption setting that the client can support.
Default setting is proprietary. --expires={NUM-DAYS | TIMESTAMP} Specifies backup expiration as number of days from today (NUM-DAYS) or an absolute TIMESTAMP.
63
avmaint COMMAND REFERENCE --flagfile=FILE Specifies a FILE containing a list of options and values that will be processed by avmaint as if they were included on the command line. Default is / usr/local/avamar/etc/usersettings.cfg. Follow output data as atomic grows. Hunt for a server storage node. Include partial backups in accounting system queries. Specifies client agent data port number. Default is 28002. Same as --acport=PORT. --mcsport=PORT Specifies data PORT that avmaint will use to communicate with the MCS. Default setting is port 28001. Sets NODE-ID for hard drive shutdowns. Turns off output to STDOUT. However, if --logfile=FILE is supplied, output will still go to log file. --nowarnings --overrideversion=NAME --rotatelogcount=NUM --rotatelogsize=MB --servbufsize=NUM --showtimestamp --sockbufsize=NUM --threadstacksize=KB Turns off warning messages. Overrides build version in XML. Specifies rotation count (NUM) of avmaint log file. Default value is 10. Specifies maximum size of avmaint log file in megabytes (MB). Specifies server TCP socket buffer size. Displays date/time stamp with STDERR messages. Specifies client TCP socket buffer size. Explicitly sets the default stack size in kilobytes (KB). Default value is zero (0), which specifies that the operating system default stack size should be used. This option added in version 3.5. --vardir=DIR Sets local cache directory. Default value is /usr/local/avamar/var. Same as --cachedir=DIR.
--node=NODE-ID --nostdout
64
Deprecated Commands
Beginning with the noted release, use of these commands is officially discouraged in favor of the better alternative. admin IMPORTANT: Deprecated in version 2.0.2 in favor of the avmaint access command (page 69). Sets server access mode to full privileges for server and root but read-only for other users (mhpu+mhpu+0000). hfscreatetime IMPORTANT: Deprecated in version 2.0.2. Returns hfscheck creation time. nobackups IMPORTANT: Deprecated in version 2.0.2 in favor of the avmaint access command (page 69). Sets server to disallow backups. offline IMPORTANT: Deprecated in version 2.0.2 in favor of the avmaint access command (page 69). Puts a stripe offline. online IMPORTANT: Deprecated in version 2.0.2 in favor of the avmaint access command (page 69). Puts a stripe back online. perfstatus [NODE-ID] IMPORTANT: Deprecated in version 4.1 in favor of the avmaint perf command (page 124). Returns detailed operational data for all server nodes or specified nodes. If one or more NODE-IDs are supplied, then detailed operational data is returned for those nodes only; otherwise, data for all nodes is returned. readmigrate IMPORTANT: Deprecated in version 2.0.2 in favor of the avmaint access command (page 69). Sets server to allow only migrate writes. readonly IMPORTANT: Deprecated in version 2.0.2 in favor of the avmaint access command (page 69). Sets server access mode to read-only for server, root and other users (0000+0000+0000). setmaster NUM IMPORTANT: Deprecated in version 1.2.1 in favor of the config command (page 76). Sets the primary module. Requires a valid module number (NUM) or none. writeable IMPORTANT: Deprecated in version 2.0.2 in favor of the avmaint access command (page 69). Sets server access mode to writeable for server, root and other users (mhpu+mhpu+mhpu).
65
66
avmaint COMMAND REFERENCE --datacenters=LIST IMPORTANT: Deprecated in version 2.0.2. Defines list of modules (datacenters) to shutdown. --disks=LIST IMPORTANT: Deprecated in version 2.0.2. Defines list of hard drives to shutdown on node. --helpx IMPORTANT: Deprecated in version 2.0.2. Prints help including extended options. --hostname=NAME IMPORTANT: Designated as obsolete in version 2.0.2. Sets name of the client machine. --ignoreconfig IMPORTANT: Deprecated in version 2.0.2. Does not read any configuration files while parsing command-line options. Same as --noconfig. --memman=NAME IMPORTANT: Deprecated in version 2.0.2. Sets memory debugging mode. --mydc IMPORTANT: Deprecated in version 2.0.2. Shuts down only this module (datacenter). --nagle IMPORTANT: Deprecated in version 2.0.2. Enables TCP socket nagling. --nat IMPORTANT: Deprecated in version 2.0.2. Sets whether the server is using Network Address Translation (NAT) between modules. --noc IMPORTANT: Deprecated in version 2.0.2. Formats output for NOC. --noconfig IMPORTANT: Deprecated in version 2.0.2. Does not read any configuration files while parsing command-line options. Same as --ignoreconfig. --nodes=NODELIST IMPORTANT: Deprecated in version 2.0.2. Defines list of nodes to shutdown. --nouserinfo IMPORTANT: Deprecated in version 2.0.2. Suppresses getting UID and GID mapping. --pizza IMPORTANT: Deprecated in version 2.0.2 in favor of --debug. Turns on debugging messages. --sessionid=NUM IMPORTANT: Designated as obsolete in version 2.0.2. Sets session ID. Same as --sid=NUM.
67
avmaint COMMAND REFERENCE --sid=NUM IMPORTANT: Designated as obsolete in version 2.0.2. Sets session ID. Same as --sessionid=NUM. --singleconn IMPORTANT: Deprecated in version 2.0.2. Makes a single connection. --stats IMPORTANT: Deprecated in version 2.0.2. Shows contents of the client message store, which is created by the Avamar server during initialization. --sysdir=PATH IMPORTANT: Deprecated in version 2.0.2. Sets directory containing Avamar server files. --syslog IMPORTANT: Deprecated in version 2.0.2. Sends log messages to the server. --uflagsdebug IMPORTANT: Deprecated in version 2.0.2. Sets debug flag parsing. --wid=WID IMPORTANT: Designated as obsolete in version 2.0.2. Sets work order ID. Same as --workorderid=WID. --workorderid=WID IMPORTANT: Designated as obsolete in version 2.0.2. Sets work order ID. Same as --wid=WID.
Notes
User Name and If your Avamar user name and password are present in your .avamar file (page Password 426), then --id=USER@AUTH and --password=PASSWORD are not required on the
command line.
Output
Output goes to STDOUT, except for errors, which go to STDERR.
68
avmaint access
The avmaint access command sets server access mode. IMPORTANT: The avmaint access command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint access commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint access {admin | normal | readmigrate | readonly | sync access} [--preconditions={available | nobackups | online} --avamaronly GLOBAL-OPTIONS
Parameters
One (and only one) of the following parameters must be supplied on each command line: admin normal readmigrate readonly sync Writing by server and root user only. Writing by everyone. Writing by server only. No writing allowed. Writing by server only.
Command Options
--preconditions= {available | nobackups | online} Specifies pre-conditions for setting server access mode. One of the following: available nobackups online Only set server access mode if all stripe groups are available. Only set server access mode if no backups are in progress. Only set server access mode if all stripes are online (none are restarting, migrating, offline); media errors are allowed.
69
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. This command added in version 2.0.2.
70
avmaint autorestart
The avmaint autorestart command controls the automatic node restart feature. If enabled, the Avamar automatic node restart feature will detect certain kinds of storage server (also known as GSAN) node failures and automatically restart a single affected node if it is determined that it is safe to do so. Various checks will be performed to ensure that the storage server (also known as GSAN) is not restarted under circumstances that could compromise system integrity: 1. All nodes must agree on the failure of a particular node before it is deemed to have actually failed. 2. Multiple node failures for any reason will inhibit the automatic node restart feature from restarting any node. 3. If the gsan process is found still to be running on an apparently failed node (this might occur if network communications are less than optimal), no action is taken. 4. A restarted gsan process will not be restarted again if it immediately fails. This restriction is removed as soon as the server reaches fullaccess mode with all stripes online. 5. If ~admin/autorestart is not present, the node is assumed to be eligible for automatic restart.
Synopsis
avmaint autorestart [--disable | --enable] [--forcestatus={up | down}[@IP-ADDR]] [--restart=IP-ADDR] [--sync] GLOBAL-OPTIONS
Command Options
--disable --enable Turns off (disables) the automatic node restart feature. Turns on (enables) the automatic node restart feature. NOTE: If the automatic node restart feature was previously disabled, and a node is in a condition such that it can be restarted, turning on the automatic node restart feature will restart that node. --forcestatus= {up | down}[@IP-ADDR] Forces perceived status of all storage server (also known as GSAN) nodes to be either up (fully operational) or down (non-functioning). Supplying the optional IP address (IP-ADDR) restricts this command to that particular storage node. --restart=IP-ADDR Forces a restart of this particular storage node.
71
avmaint autorestart COMMAND REFERENCE --sync Resynchronizes ascd to correct a situation in which the probe.out file might be changed while the server is running, which will occur if a node is decommissioned and replaced during service.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Files
~admin/autorestart On multi-node systems, if ~admin/autorestart is present on a storage node, it contains information related to the last unexpected shutdown of that node. This information is used to determine if that node is eligible to be automatically restarted. If ~admin/autorestart is not present, the node is assumed to be eligible for an automatic restart.
Example
avmaint autorestart --xmlperline=99
<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <autorestart enabled="true" state="monitor"> <gsan ipaddr="10.6.252.90" nodeid="0.0" state="up"/> <gsan ipaddr="10.6.252.91" nodeid="0.1" state="up"/> <gsan ipaddr="10.6.252.92" nodeid="0.2" state="up"/> <gsan ipaddr="10.6.252.93" nodeid="0.3" state="up"/> <gsan ipaddr="10.6.252.94" nodeid="0.4" state="up"/> <gsan ipaddr="10.6.252.95" nodeid="0.5" state="up"/> <gsan ipaddr="10.6.252.96" nodeid="0.6" state="up"/> <gsan ipaddr="10.6.252.97" nodeid="0.7" state="up"/> </autorestart>
Notes
This command added in version 4.1.
72
avmaint cancelremovebadhashes
The avmaint cancelremovebadhashes command cancels a pending find/delete bad hash operation and returns the server to a fully operation state. Refer to avmaint findbadhashes (page 91) and avamaint removebadhashes (page 131) for additional information. IMPORTANT: The avmaint cancelremovebadhashes command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint cancelremovebadhashes commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint cancelremovebadhashes --avamaronly GLOBAL-OPTIONS
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
This command added in version 4.1.
73
avmaint cat
The avmaint cat command lists contents of a persistent ATOMIC.
Synopsis
avmaint cat ATOMIC [--follow] [--maxbytes=LENGTH] GLOBAL-OPTIONS
Parameters
ATOMIC Specifies which ATOMICs contents to list.
Command Options
--follow --maxbytes=LENGTH Forces following object tail until interrupted. Specifies maximum byte LENGTH.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
This command added in version 2.0.2.
74
avmaint checkpoint
The avmaint checkpoint command starts a checkpoint operation. IMPORTANT: The avmaint checkpoint command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint checkpoint commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint checkpoint [--keepmin] [--wait] [--waittime=SEC] --avamaronly GLOBAL-OPTIONS
Command Options
--keepmin Enables the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. This option added in version 4.0. --wait --waittime=SEC Forces return of a new checkpoint ID only after successful checkpoint completion. Specifies number of seconds (SEC) to wait before actually initiating the checkpoint in order to allow clients to terminate gracefully. Default value is 600 seconds.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. Beginning with version 3.0.1, this command now returns a new checkpoint ID immediately (before the checkpoint completes).
75
avmaint config
The avmaint config command configures various internal characteristics of the server. If avmaint config is run without parameters, it returns a current list of server configuration settings. IMPORTANT: The avmaint config command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint config commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint config {asynccrunching=BOOL | autorepairmax=NUM | balancemin=NUM | cacheupdateinterval=SEC | chunkhashcheck=BOOL | commtimeout=NUM | compressed | cphfschecked=NUM | cpkeepmin=BOOL | cpmostrecent=NUM | disknocp=PERCENT | disknocreate=PERCENT | disknoflush=PERCENT | disknogc=PERCENT | diskreadonly=PERCENT | diskwarning=PERCENT | failrefcheck=BOOL | hfsport=PORT | hintcachemax=NUM | indexelements=NUM | indexsplitcount=PERCENT | indexsplitspread=NUM | iolimitpercent=PERCENT | lmaddr=IP-ADDR | lmport=PORT | masterdc=NUM | maxcompdatastripe=SIZE | maxconn=NUM | maxconninactive=SEC | maxdatastripe=SIZE | maxlogfiles=NUM | maxlogsize=MB | maxrequests=NUM | maxrwatomdatastripe=SIZE | maxrwcompdatastripe=SIZE | minstripestowrite=NUM | paritygroups=Nx,Fy | perfbufsize=NUM | perfhistory=DAYS | perfinterval=SEC | refcheck=BOOL | sslport=PORT | vmstatdelay=NUM | vmstatprofile{PROFILE-NAME | NUM)} --avamaronly GLOBAL-OPTIONS
Parameters
One (and only one) of the following parameters must be supplied on each command line: asynccrunching=BOOL Enables or disables asynchronous stripe crunching. To enable this feature, set BOOL to true (caseinsensitive) or one (1). To disable this feature, set BOOL to false (caseinsensitive) or zero (0). Asynchronous stripe crunching is always enabled following a restart. This parameter added in version 3.6.
76
avmaint config COMMAND REFERENCE autorepairmax=NUM Specifies maximum number (NUM) of stripes with errors that will be fixed by hfscheck. Default value is 8. The maximum value is 1024. If the value is 0, no automatic repair will be performed. If there are more stripes with errors than the current autorepairmax setting, then no repairs will be performed (errors will be generated for all stripes). This parameter added in version 3.7.1. balancemin=NUM Configures server load balancing. When this value is zero, no balancing is performed. When this value is one, only index stripes are balanced. For values between 2 and 99, this is a threshold percentage that will trigger load balancing. For example, a setting of 2 will trigger rebalancing of a storage node if it is 0.2 percent less dense than the average density of all other storage nodes; a setting of 34 will trigger rebalancing of a storage node if it is 3.4 percent less dense than the average density of all other storage nodes. cacheupdateinterval=SEC Control the frequency of the cachebeat notifications, which notifies the system that stripes have changed so that the writelog header can be updated. Default setting is 300 seconds (5 minutes). Allowable values are 30 to 65535 seconds. Specifying a zero value disables the cachebeat. This parameter added in version 3.7. chunkhashcheck=BOOL Enables or disables additional hash validation in datastripe getchunk. Default value is true. To enable this feature, set BOOL to true (caseinsensitive) or one (1). To disable this feature, set BOOL to false (caseinsensitive) or zero (0). commtimeout=NUM Specifies number (NUM) of seconds used for internal server communication time-outs. Default value is 30. This parameter added in version 3.5. compressed cphfschecked=NUM Enables or disables backup compression. Default value is true (compress backups). Specifies minimum number (NUM) of validated (HFS-checked) checkpoints that must be retained in the system at all times. Default value is 1. This parameter added in version 3.0.1.
77
avmaint config COMMAND REFERENCE cpkeepmin=BOOL Persistently enables or disables the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. To enable this feature, set BOOL to true (caseinsensitive) or one (1). This is the default setting. To disable this feature, set BOOL to false (caseinsensitive) or zero (0). This parameter added in version 4.0. cpmostrecent=NUM Specifies minimum number (NUM) of most-recent checkpoints that must be retained in the system at all times. Default value is 2. This parameter added in version 3.0.1. disknocp=PERCENT Does not create new checkpoints after this percentage (PERCENT) of full server storage capacity is reached. Does not create new stripes if more than this percentage (PERCENT) of full server storage capacity is reached. Does not flush writelogs to stripes if more than this percentage (PERCENT) of full server storage capacity is reached. Does not perform garbage collect after this percentage (PERCENT) of full server storage capacity is reached. Sets percentage (PERCENT) of full server storage capacity that will trigger conversion of the server from a fully writable condition to read-only. Default value is 65%. Sets percentage (PERCENT) of full server storage capacity that will trigger a warning to the server administrator that the server is becoming full. If refcheck is set true, failrefcheck controls whether a refcheck error causes the add_hash_data to fail. When debugging, it is sometimes useful to allow the operation to succeed so that you can see where the hash occurred in the backup. To enable this feature, set BOOL to true (caseinsensitive) or one (1). This is the default setting. To disable this feature, set BOOL to false (case-insensitive) or zero (0).
disknocreate=PERCENT
disknoflush=PERCENT
disknogc=PERCENT
diskreadonly=PERCENT
diskwarning=PERCENT
failrefcheck=BOOL
78
avmaint config COMMAND REFERENCE hfsport=PORT Specifies port ID (PORT) for client communications using Avamar proprietary encryption. Default value is 27000. IMPORTANT: Changing this port assignment will cause the Avamar server to become inaccessible to any Avamar clients that have not similarly changed to the new port ID. This parameter added in version 3.5. hintcachemax=NUM Specifies number of entries in the hint cache. If set to zero, no hint caching is performed. Otherwise, this value specifies the maximum number (NUM) of entries in the hint cache. This parameter added in version 3.7. indexelements=NUM indexsplitcount=PERCENT Sets index stripe size to this number (NUM) of elements. Sets maximum percentage (PERCENT) of index elements that can exist in an index stripe before it is programmatically split. Default value is 85%. indexsplitspread=NUM iolimitpercent=PERCENT Sets variation in indexsplitcount across stripes. Controls the number of simultaneous I/O requests issued by the data server (also known as GSAN) to each disk in the system. This parameter interacts with the operating system nr_requests threshold as follows: limit = nr_requests * (iolimitpercent / 100) This parameter added in version 4.0. lmaddr=IP-ADDR Specifies IP address (IP-ADDR) of external authentication manager or zero (do not use an external authentication manager). Default value is zero. IMPORTANT: Changing this setting can cause external authentication to fail if the assigned port is not correct. This parameter added in version 3.5. lmport=PORT Specifies port ID (PORT) of external authentication manager. Default value is 700. IMPORTANT: Changing this port assignment can cause external authentication to fail if the assigned port is not correct. This parameter added in version 3.5. masterdc=NUM Sets master module to N. Default value is 0.
79
avmaint config COMMAND REFERENCE maxcompdatastripe=SIZE maxconn=NUM maxconninactive=SEC Sets maximum composite data stripe SIZE to this number of elements. Sets maximum number (NUM) of concurrent client connections for each node. Sets client connection time-out period in seconds (SEC). This is the maximum amount of time that the server will wait for a client to resume communication after network connectivity is lost. If client communication is not resumed within this period of time, the server terminates the client connection. Sets maximum atom data stripe SIZE to this number of elements. Sets maximum number of gsan.log files to retain. Default value is 25 MB. NOTE: In multi-node servers, this setting applies equally to all storage nodes on the server (you cannot configure gsan.log file size on a node-by-node basis). maxlogsize=MB Sets maximum gsan.log file size in megabytes (MB). Default value is 25 MB. NOTE: In multi-node servers, this setting applies equally to all storage nodes on the server (you cannot configure gsan.log file size on a node-by-node basis). maxrequests=NUM maxrwatomdatastripe=SIZE maxrwcompdatastripe=SIZE minstripestowrite=NUM Sets maximum number (NUM) of concurrent client requests per node. Default value is 150. Sets maximum read-write (rw) atomic data stripe SIZE to this number of elements. Sets maximum rw composite data stripe SIZE to this number of elements. Sets minimum number (NUM) of stripes per family that must be online in order to write.
maxdatastripe=SIZE maxlogfiles=NUM
80
avmaint config COMMAND REFERENCE paritygroups=Nx,Fy Sets parity description by specifying the sizes of near and far parity groups, where Nx is the maximum size of near parity groups and Fy is the maximum size of the first far parity group. Larger parity groups improve storage efficiency at the expense of reconstruction efficiency. In other words, large parity groups use disk space more efficiently, but take more time and more disk seeks to backup or restore data when a node or module is down. In multi-node servers, the maximum near parity setting is always the number of storage nodes in a module. Far parity groups will never grow larger than the number of modules minus one. Default value is N8,F4. perfbufsize=NUM Specifies read performance buffer size according to the following formula: 128KB * 2NUM NUM can be any whole integer between 0-4, which results in the following buffer sizes: 0 1 2 3 4 128KB buffer. 512KB buffer. 1024KB buffer. 2048KB buffer. 4096KB buffer.
Default setting is 3 (2048KB buffer). This option added in version 4.0. perfhistory=DAYS Specifies number of DAYS of read performance history to retain. Valid values are 2 thru 255. Default setting is 10 days. This option added in version 4.0. perfinterval=SEC Specifies number of seconds (SEC) between read performance checks. Default setting is 300 seconds. A setting of zero disables this feature. This option added in version 4.0. refcheck=BOOL Enables or disables check composite/directory references on add hash data. To enable this feature, set BOOL to true (caseinsensitive) or one (1). This is the default setting. To disable this feature, set BOOL to false (case-insensitive) or zero (0).
81
avmaint config COMMAND REFERENCE sslport=PORT Specifies port ID (PORT) for client communications by way of SSL. Default value is 29000. IMPORTANT: Changing this port assignment will cause the Avamar server to become inaccessible to any Avamar clients that have not similarly changed to the new port ID. This parameter added in version 3.5. vmstatdelay=NUM Specifies number (NUM) of seconds between vmstat log file write operations. Default value is 60 seconds. This parameter added in version 3.5. vmstatprofile= {PROFILE-NAME | NUM) Controls selection of kernel information written to the log file. This mechanism is similar to the Unix vmstat command, where PROFILE-NAME is a profile string name or NUM is the numerical sum of all data profiles that are to be used. Valid values are one or more of the following string or numeric profile values: {active | 0x00000400} Active profile. Tracks number of datastripes and composite datastripes that are currently active (ready to be written to). Chunk profile. Tracks number of chunks and bytes that have been added, queued for deletion by garbage collection or deleted. It also provides rates of change for these statistics. Client profile. Tracks number of client backups and restores in progress and causes special clientdone messages to be written to the server logs following a successful client operation in order to provide data transfer statistics.
{chunk | 0x00000100}
{client | 0x00000200}
82
avmaint config COMMAND REFERENCE vmstatprofile= {PROFILE-NAME | NUM) (continued) {disk | 0x00000008} Disk usage profile. Tracks disk usage and I/ O. If the stripe profile is also enabled, stripe information per disk is integrated in to the results. Event profile. Tracks number of checkpoint, stripe creation, datastripe crunch, garbage collection, indexstripe split and stripe sync operations in progress. It also tracks total number of these operations that have occurred and whether an hfscheck is in progress. Flush profile. Tracks stripe flush activity (the number of stripes flushed, the total number of seconds spent flushing and the average number of seconds per flush). Load average profile. Provides current node load averages as returned by Linux. Network profile. Tracks network packet I/O for TCP and UDP. Node state profile. Provides current state of a node (online status, access mode, run level, stripe count, freezing status, connectivity stability, command dispatcher and scheduler suspension states). Memory pool profile. Tracks state of primary memory pool.
{event | 0x00000800}
{flush | 0x00001000}
{load | 0x00000020}
{net | 0x00000002}
{node | 0x00000040}
{pool | 0x00000080}
83
avmaint config COMMAND REFERENCE vmstatprofile= {PROFILE-NAME | NUM) (continued) {rawnet | 0x00000004} Raw network profile. Tracks raw network I/O data including bytes in and out, packets in and out, and errors. Standard profile. Displays the same information as the Linux vmstat command. Stripe info profile. Provides merged run-time stripe info (opens, closes, I/O operations summaries, and so forth). Stripe count profile. Tracks number of stripes of each kind and parity existence on each node. Uniform profile. This is a specialpurpose formatting argument, that when specified, formats vmstat log messages (produced by other profiles) in manner that can be more easily processed by a script. Each uniform log message begins with the profile name and is followed by one or more NAME=VALUE pairs.
{std | 0x00000001}
{stripe | 0x00000010}
{stripecount | 0x00002000}
{uniform | 0x00004000}
84
avmaint config COMMAND REFERENCE vmstatprofile= {PROFILE-NAME | NUM) (continued) Multiple string profile names can be supplied if separated by commas (for example, vmstatprofile=chunk,stripecount). Multiple numeric profiles can be specified as a single exclusive ORd numeric value (for example, vmstatprofile=0x00002100 is equivalent to vmstatprofile=chunk,stripecount). This parameter added in version 3.5.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML.
85
avmaint conversion
The avmaint conversion command converts older format data stripes to latest version. IMPORTANT: The avmaint conversion command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint conversion commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint conversion [--count=NUM] [--maxtime=SEC] --avamaronly GLOBAL-OPTIONS
Command Options
--count=NUM Specifies maximum number (NUM) of stripes that will be converted on each node. Default value is 50 stripes per node. A setting of zero specifies that an unlimited number of stripes be converted (convert all stripes now). NOTE: This is an absolute maximum setting. The actual number of stripes converted might be less, especially if the --maxtime setting is a very short duration. --maxtime=SEC Specifies maximum amount of time that avmaint conversion will be allowed to run. Default value is 900 seconds (15 minutes). A setting of zero specifies an unlimited amount of time (run until all stripes are converted).
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. This command added in version 3.5.
86
avmaint crunch
The avmaint crunch command controls stripe crunching. IMPORTANT: The avmaint crunch command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint crunch commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint crunch {limit LIMIT | status | reset | rollover | resetandrollover | factoryreset} --avamaronly GLOBAL-OPTIONS
Parameters
One (and only one) of the following parameters must be supplied on each command line: limit LIMIT Changes the daily amount of asynchronous crunching in accordance withe the supplied LIMIT value, where LIMIT is a singed integer representing some change in the perecentage of daily asynchronous crunching that should take place. If the value is positive, then the new goal will be G + (G * LIMIT%). If the value is negative, then the new goal will be G - (G * LIMIT%), but never less than zero. The goal adjustment is only used until the crunch day rolls over, which occurs at midnight on the storage node, when an avmaint crunch rollover or avmaint crunch resetandrollover command is received, or when the server is restarted status reset rollover resetandrollover factoryreset Returns crunching status only; no changes made. Sets counters to 0. Rolls over day timer to current time. Equivalent to performing reset, then rollover actions. Sets day timer to factory default (midnight gmt).
Command Options
--avamaronly Enables commands and options that are normally intended for use by expert users only. Contact EMC Technical Support for additional information. IMPORTANT: --avamaronly must be supplied in order for this command to work.
87
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. This command added in version 3.6. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
88
avmaint decommission
The avmaint decommission command decommissions a storage node or an individual disk within a storage node. IMPORTANT: The avmaint decommission command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint decommission commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint decommission {NODE-ID | --disknum=DISK} --avamaronly
Parameters
One (and only one) of the following parameters must be supplied on each command line: NODE-ID --disknum=DISK If a NODE-ID (for example, 0.1, 0.2) is supplied, that node is decommissioned. If --disknum=DISK is supplied, that DISK is decommissioned.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --expert IMPORTANT: Deprecated in version 2.0.2. Performs command even if server is unavailable.
Notes
You must include the --avamaronly option. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
89
avmaint ducp
The avmaint ducp command lists hard drive space consumed by a specific checkpoint (CP-ID).
Synopsis
avmaint ducp CP-ID GLOBAL-OPTIONS
Parameters
CP-ID Specifies which checkpoint to list.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
90
avmaint findbadhashes
The avmaint findbadhashes command returns a list of hashes and backups that are suspected to be bad, as well as a verification token, which must be supplied to removebadhashes (page 131) in order to actually remove bad hashes from the server. IMPORTANT: The avmaint findbadhashes command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint findbadhashes commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
IMPORTANT: In order to ensure data integrity, avamaint findbadhashes places the server in read-migrate mode, which terminates any running backups and prevents any new backup, restore or replication operations from taking place until such time as avamaint removebadhashes (page 131) completes. If you must revert to fully operational server state before running avamaint removebadhashes (page 131), use avamaint cancelremovebadhashes (page 73) to un-do this entire find/delete bad hash operation.
Synopsis
avmaint findbadhashes [H] HASH --avamaronly GLOBAL-OPTIONS
Parameters
H HASH HFS check hash. 40-character hex HASH value, which is typically obtained from avmaint hfscheck avmaint hfscheck (page 100) error messages.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
91
Output
Output is XML in the following format:
<?XML VERSION="1.0" ENCODING="UTF-8" STANDALONE="YES"?> <FINDBADHASHESLIST PASSCOUNT="15" CREATED="1213311581" VERIFICATION-TOKEN="VTOKENC7DC5D67658D24785F3607803E3B130FEFD8B432"> <HASHLIST> <HASH VALUE="H253F32A5DD143D424BCCD352DED54EF311100B38"> <BACKUP CLIENT="CLIENTS/MYCLIENT.EXAMPLE.COM" LABELNUM="1" CREATED="1208471840"> <DIRPATH PATH="C:/TEMP/FILE-1.TXT"/> </BACKUP> </HASH> <HASH VALUE="HDA140B39BC2F89EEC0FE076FFFC468064C21CBA0"> <BACKUP CLIENT="NONE" LABELNUM="" CREATED=""> <DIRPATH PATH="/DATA02/FILE-2.TXT"/> <DIRPATH PATH="/DATA02/FILE-3.TXT"/> <DIRPATH PATH="/DATA02/FILE-4.TXT"/> </BACKUP> </HASH> </HASHLIST> </FINDBADHASHESLIST>
Notes
This command added in version 4.1.
92
avmaint garbagecollect
The avmaint garbagecollect command starts a garbage collect operation. IMPORTANT: The avmaint garbagecollect command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint garbagecollect commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint garbagecollect [--gccount=NUM] [--kill=NUM] [--limitadjust={+ | -}LIMIT] [--maxpass=NUM] [--maxtime=SEC] [--refcheck] [--throttlelevel=PERCENT] [--usehistory] --avamaronly GLOBAL-OPTIONS
Command Options
--gccount=NUM --kill=NUM Specifies number (NUM) of index stripes per node to garbage collect. Default value is 16. Forcibly terminates (kills) backups after waiting this number (NUM) of seconds. Default setting is zero (0), which disables this feature (that is, backups are not forcibly terminated). This option added in version 4.0. --limitadjust= {+ | -}LIMIT Used with --usehistory to increase or decrease computed garbage collect quota by LIMIT percent. Default setting is zero (0). For example, if history computes that NN GB should be garbage collected, then using --limitadjust=+50 increases that amount to NN + (NN*50%) GB. Alternatively, supplying --limitadjust=-25 decreases that amount to NN - (NN*25%) GB. This option added in version 4.1. --maxpass=NUM --maxtime=SEC Specifies maximum number (NUM) of garbage collection passes. Default value is 100 passes. Specifies maximum number of seconds (SEC) garbage collect process will be allowed to run. Default value is 3600 seconds (1 hour). Forces check of composite references during garbage collect or HFS check, respectively.
--refcheck
93
avmaint garbagecollect COMMAND REFERENCE --throttlelevel=PERCENT Specifies throttling percentage (PERCENT). This value reduces the garbage collect operations system resource (CPU, disk, bandwidth, and so forth) usage by this percentage. Default setting is zero (that is, no throttling), which means that the garbage collect operations system resource usage is potentially 100%). This option added in version 4.0. --usehistory Enables or disables garbage collect history feature, which limits the amount of garbage collection performed to an amount that will free up at least as much space as has been used by the daily average of all backups over the last several days. Default is false. This option added in version 4.1.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
94
avmaint gcstatus
The avmaint gcstatus command returns status for an on-going or the most recently completed garbage collect process.
Synopsis
avmaint gcstatus [--full] GLOBAL-OPTIONS
Command Options
--full Returns full listings.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Beginning with version 3.0.1, supplying --full returns information on each node as well as the global summary. Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
95
avmaint gendata
The avmaint gendata command generates data chunks for testing purposes. IMPORTANT: The avmaint gendata command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint gendata commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint gendata [--localindex] [--pending=NUM] [--starttime=NUM] [--totalmb=MB] --avamaronly GLOBAL-OPTIONS
Command Options
--localindex --pending=NUM --starttime=NUM --totalmb=MB Add to local index stripes. Specifies number (NUM) of simultaneous write requests. Default setting is 10. Specifies time to start test. Total megabytes (MB) of data to write. Default setting is 1.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option.
96
avmaint getclientmsgs
The avmaint getclientmsgs command returns contents of the client message store. Messages provide session information about all avtar sessions that have run or are being run on the Avamar server.
Synopsis
avmaint getclientmsgs [--startoffset=NUM] GLOBAL-OPTIONS
Command Options
--startoffset=NUM Specifies starting offset. Default value is 0. You might need to repeatedly invoke this command until the return EOF is 1 in order to view all of the messages.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Support for non-XML output is likely to be discontinued in a future release. You should include --xml and the --xmlperline=NUM global option to specify XML output and to control the number of XML attributes per line, respectively.
97
avmaint geterrors
The avmaint geterrors command returns errors from a specified NODE-ID.
Synopsis
avmaint geterrors NODE-ID [--alternate] [--duptime=SEC] [--errfilter={ERROR | INFO | WARN}] [--startoffset=NUM] GLOBAL-OPTIONS
Parameters
NODE-ID Return errors from the specified NODE-ID, which can be obtained by using the nodelist command (page 122).
Command Options
--alternate --duptime=SEC Specifies that errors be recovered from the hfscheck log file rather than the server log. Returns errors within the specified period of time (in seconds) since the server was started or restarted. Default value is 60 seconds. Specifies which types of errors (ERROR, INFO or WARN) to return. You can specify more than one error type, but multiple error types must be separated by commas. For example, supplying --errfilter=ERROR,WARN will return errors and warnings. --startoffset=NUM Defines a starting range (line number) inside the log file. Default value is zero.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
98
avmaint getrefby
The avmaint getrefby command returns all hashes of a particular type that reference this HASH. IMPORTANT: The avmaint getrefby command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint getrefby commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint getrefby {H | R | U} HASH --avamaronly GLOBAL-OPTIONS
Parameters
H R U HASH HFS check hash. Persistent store hash. User account hash. 40-character hex HASH value.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option.
99
avmaint hfscheck
The avmaint hfscheck command initiates an HFS check operation. It returns control to the user immediately after the HFS check process (cgsan) has initiated the check. IMPORTANT: The avmaint hfscheck command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint hfscheck commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it. If avmaint hfscheck is run without options, an HFS check is performed on the most recent unvalidated checkpoint file. To return status for a currently-running or the most recently completed HFS check, use avmaint hfscheckstatus (page 105). To stop a currently-running HFS, use avmaint hfscheckstop (page 60) or hfscheck_kill (page 294). Refer to Checkpoint Validation (HFS Checks) (page 15) for additional detailed technical information about the HFS check feature.
Synopsis
avmaint hfscheck --checks=DESCRIPTOR [--checknode=MODULE.NODE] [--checkpoint=CP-ID] [--concurrentdatastripes=NUM] [--concurrentindexstripes=NUM] [--concurrentparitystripes=NUM] [--cxorsets=NUM] [--deadline=MIN] [--gccount] [--keep] [--keepmin] [--resetpredictor] [--suspend] [--throttlelevel=NUM] --avamaronly GLOBAL-OPTIONS
100
Command Options
--checks=DESCRIPTOR DESCRIPTOR must be a valid enabling descriptor string that describes which checks are to be performed on which specific stripe classes. Stripe classes can be any or all of the following: h p u HFS stripes. Persistent stripes. Accounting stripes.
Checks can be any or all of the following: a c p r Atomic data sweeps. Composite data sweeps. Parity checks. Reference checking.
For example, this descriptor: hpu+acpr Is equivalent to the default behavior (perform all checks on all stripe classes). You can also constrain an HFS check to only process a percentage of the total stripes by prefixing a desired processing percentage (as an integer between 1 and 99). For example: 50:hpu Default value is 100 (process 100% of eligible stripes). Refer to HFS Check Enabling Descriptor (page 17) for additional information about enabling descriptor syntax. --checknode=MODULE.NODE Enables single-node HFS check (page 21) on this specific MODULE.NODE. This option added in version 3.6. --checkpoint=CP-ID Specifies which checkpoint (CP-ID) should be processed. If not supplied, HFS check is performed on the most recent unvalidated checkpoint file. This option added in version 3.5.
101
avmaint hfscheck COMMAND REFERENCE --concurrentdatastripes=NUM Specifies maximum number of data stripes that are simultaneously processed on a single node during a index sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5. --concurrentindexstripes=NUM Specifies maximum number of index stripes that are simultaneously processed on a single node during an index sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5. --concurrentparitystripes=NUM Specifies maximum number of parity stripes that are simultaneously processed on a single node during a parity sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5. --cxorsets=NUM Specifies number of column XORs sets to use for parity checking during an HFS check. Default value is 0 (use the data server default, which is currently 3). This option added in version 3.5. --deadline=MIN Specifies HFS check deadline (page 21) in minutes (MIN). Default value is zero (no deadline). This option added in version 3.6. --gccount Specifies maximum number of index stripes that are simultaneously processed on a single node during the refcheck phase. This option added in version 3.5. --keep Specifies that temporary working directories and files should be retained (kept) after HFS check completes. Default value is to delete all temporary working files. This option added in version 3.5. --keepmin Enables the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. This option added in version 4.0.
102
avmaint hfscheck COMMAND REFERENCE --resetpredictor --suspend Forces a re-initialization of predictor logic. Specifies that dispatchers should be suspended during HFS check program start up (at the same time that server access mode is set to read-only). This option added in version 3.6. --throttlelevel=NUM Specifies the amount of throttled resources (measured as a percentage of CPU capacity) that can be allocated to other tasks. For example, a setting of 20 begins throttling once CPU usage exceeds 80%. Values in excess of 100 are treated as 100% (maximum throttling); a value of zero specifies that no throttling is to be performed. If a deadline is specified, the throttle level indicates a maximum allowable throttle level, and the throttle level starts at half this value. Default value is 20 if no deadline is specified; 40 if deadline is specified. Refer to HFS Check Throttling (page 19) for additional information about throttle levels and deadlines. This option added in version 3.5.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --concurrentstripes=NUM IMPORTANT: Deprecated in version 3.5 in favor of --concurrentdatastripes, --concurrentindexstripes and --concurrentparitystripes. Specifies number (NUM) of HFS check index and parity stripes per node to check concurrently. Default value is 1, which is equivalent to: --concurrentindexstripes=1 --concurrentdatastripes=1 --concurrentparitystripes=1
103
avmaint hfscheck COMMAND REFERENCE --paritymode=MODE IMPORTANT: Deprecated in version 3.5 in favor of --checks. Specifies how parity stripes should be checked during an hfscheck. MODE must be one of the following: columnxor none Default method of parity checking is performed. No parity checking is performed.
Default value is columnxor. --refcheck IMPORTANT: Deprecated in version 3.5 in favor of --checks. Forces check of composite references during garbage collect or HFS check, respectively.
Notes
You must include the --avamaronly option. The command returns after immediately after the hfscheck process has started. This can take several minutes. However, the actual HFS check operation can take several hours. Any throttling applied to the HFS check operation will increase the amount of time it takes to complete. An HFS check consumes extensive amounts of server resources. In order to reduce contention with normal server operation, an HFS check can be throttled. Refer to HFS Check Throttling (page 19) for additional information. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
Output
On return, the following information is output in XML format indicating a successful initiation of the HFS check:
<?xml version ="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE hfscheck> <hfscheck checkpoint="cp.20051215003909" status="hfscheck" start-time="1134607161" end-time="0" elapsed-time="246" check-start-time="1134607407"/>
Refer to avmaint hfscheckstatus (page 105) for a complete description of these attributes.
104
avmaint hfscheckstatus
The avmaint hfscheckstatus command returns the status of any currentlyrunning HFS check or the last completed HFS check. Completed information is retained until the data server is restarted, at which point a new HFS check is required in order to update this information.
Synopsis
avmaint hfscheckstatus [CP-ID] GLOBAL-OPTIONS
Command Options
CP-ID If checkpoint ID (CP-ID) is supplied, avmaint hfscheckstatus returns status for that specific HFS check. If checkpoint ID (CP-ID) is not supplied, avmaint hfscheckstatus returns status for the most recently completed HFS check. This option added in version 4.0.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
Output
Status is output in XML in the following format:
<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <hfscheckstatus nodes-queried="4" nodes-replied="4" nodes-total="4" generation-time="1218057050" checkpoint="" start-time="0" end-time="0" elapsed-time="0" check-start-time="0" check-end-time="0" status="idle" type="full" checks="0:"> <hfscheckerrors/> </hfscheckstatus>
105
avmaint hfscheckstatus COMMAND REFERENCE The following table describes each avmaint hfscheckstatus XML attribute:
ATTRIBUTE DESCRIPTION
Number of nodes that were sent a status request. Number of nodes that responded to status request. Total number of server nodes. Time, expressed as seconds of epoch, that the avmaint hfscheckstatus command was issued. Name of checkpoint that was last checked. Time, expressed as seconds of epoch, that the initiate HFS check command was issued. Time, expressed as seconds of epoch, that the HFS check operation ended. If status is for a completed HFS check operation, this is the total elapsed time in seconds that the HFS check operation consumed. If status is for a currently-running HFS check operation, this is the current amount time in seconds that has elapsed.
check-start-time
Time, expressed as seconds of epoch, that the HFS check starts processing on the newly created HFS check process (cgsan). Time, expressed as seconds of epoch, that the HFS check ended.
check-end-time
106
status
One of the following status codes: commit completed error Performing commit operation to generate HFS check work directory. HFS check has successfully completed without errors. HFS check has completed but with an error. NOTE: This generally indicates a procedural error rather than an actual defect in the HFS. hfscheck idle HFS check operation has started on cgsan but not yet completed. No HFS check has been performed since the last data server start or restart. Performing rollback operation on checkpoint directory. Forking and executing cgsan. HFS check has completed but final clean up is occurring. Awaiting cgsan startup. This occurs once cgsan reaches full access mode.
type
Check type (page 22). One of the following: full reduced partial All checks are performed on all stripes. Checkpoint was taken on N nodes but checked on N-1 nodes. All other kinds of checks.
checks phase
If type is not full, this is the check enabling descriptor (page 17). The element is only present if status=hfscheck. One of the following: datasweep indexsweep paritysweep refcheck This element added in version 3.6.
107
minutes-to-completion
The element is only present if status=hfscheck. Provides and estimated number of minutes to completion of currently-running HFS check. This element added in version 3.6.
OK if HFS check successfully completed. Otherwise, an error code is provided. Estimated number of stripes that remain to be checked. Number of stripes that were actually checked. It might be less than that estimated if certain operations (parity checking, for example) are not performed. While an HFS check is currentlyrunning, this gives the current checked figure. Number of checking errors that have been detected.
errors
The actual errors generated can be obtained with the avmaint geterrors --alternate command (page 98). Refer to Common hfscheck Error Codes (page 16) for additional information.
108
avmaint hfscheckthrottle
The avmaint hfscheckthrottle command dynamically modifies existing throttling values for a currently-running HFS check operation. IMPORTANT: The avmaint hfscheckthrottle command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint hfscheckthrottle commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it. Some important limitations apply to modifying HFS check throttling values once the operation has started. For example, once the index sweep pass has terminated, the corresponding concurrency parameter is no longer used and therefore changing the concurrent index stripes setting will have no effect. However, the lengthy passes such as the parity or data sweeps can be changed during their course though this might take a few seconds or minutes to become operative.
Synopsis
avmaint hfscheckthrottle [--concurrentdatastripes=NUM] [--concurrentindexstripes=NUM] [--concurrentparitystripes=NUM] [--gccount] [--throttlelevel=NUM] --avamaronly GLOBAL-OPTIONS
Command Options
--concurrentdatastripes=NUM Specifies maximum number of data stripes that are simultaneously processed on a single node during a index sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5. --concurrentindexstripes=NUM Specifies maximum number of index stripes that are simultaneously processed on a single node during an index sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5. --concurrentparitystripes=NUM Specifies maximum number of parity stripes that are simultaneously processed on a single node during a parity sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5.
109
avmaint hfscheckthrottle COMMAND REFERENCE --gccount Specifies maximum number of index stripes that are simultaneously processed on a single node during the refcheck phase. This option added in version 3.5. --throttlelevel=NUM Specifies the amount of throttled resources (measured as a percentage of CPU capacity) that can be allocated to other tasks. For example, a setting of 20 begins throttling once CPU usage exceeds 80%. Values in excess of 100 are treated as 100% (maximum throttling); a value of zero specifies that no throttling is to be performed. If a deadline is specified, the throttle level indicates a maximum allowable throttle level, and the throttle level starts at half this value. Default value is 20 if no deadline is specified; 40 if deadline is specified. Refer to HFS Check Throttling (page 19) for additional information about throttle levels and deadlines. This option added in version 3.5.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --concurrentstripes=NUM IMPORTANT: Deprecated in version 3.5 in favor of --concurrentdatastripes, --concurrentindexstripes and --concurrentparitystripes. Specifies number (NUM) of HFS check index and parity stripes per node to check concurrently. Default value is 1, which is equivalent to: --concurrentindexstripes=1 --concurrentdatastripes=1 --concurrentparitystripes=1
Notes
You must include the --avamaronly option. This command added in version 3.5. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
110
avmaint infomessage
The avmaint infomessage command writes an informational text message to the server log. IMPORTANT: The avmaint infomessage command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint infomessage commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint infomessage TEXT [--errcode=NUM] [--errkind={error | info | warning}] --avamaronly GLOBAL-OPTIONS
Parameters
TEXT Informational TEXT message that will be written to the server log.
Command Options
--errcode=NUM --errkind= {error | info | warning} Specifies an error code number (NUM) for this message. Specifies the type of error message. One of the following: error info warning Designate this message as an error messages. Designate this message as an informational messages. Designate this message as an warning messages.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option.
111
avmaint kill
The avmaint kill command stops (kills) this client SESSION-ID.
Synopsis
avmaint kill SESSION-ID [--waittime=SEC] GLOBAL-OPTIONS
Parameters
SESSION-ID Specifies which client SESSION-ID to stop (kill).
Command Options
--waittime=SEC Specifies number of seconds (SEC) to wait before actually terminating client sessions in order to allow clients to terminate gracefully. Default value is 600 seconds.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
This command added in version 2.0.2.
112
avmaint logscan
The avmaint logscan command enables enhanced server log scanning, which will detect and report certain kinds of hardware failures in addition to Avamar software errors. IMPORTANT: The avmaint logscan command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint logscan commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it. avmaint logscan uses an input file, which contains a list of log file descriptors (that is, log files that will be scanned) and matching patterns for each log file that will be evaluated. Each log file descriptor must include at least one pattern used to match lines in the corresponding log file. Each pattern is given a chance to evaluate each new entry in the log file. Patterns are tested in the order specified by the optional order attribute, or in some undefined order if no order attributes are provided. Once a pattern is matched, no further testing is performed. All patterns without the optional order attribute will be tested after any patterns that specify an order attribute. An optional boolean enabled attribute indicates whether the pattern should be used. If the enabled attribute is not present, the pattern will be tested.
Synopsis
avmaint logscan [FILE] --avamaronly GLOBAL-OPTIONS
Parameters
FILE If supplied, specifies new enhanced server log scanning parameters information and return old information. FILE must conform to the format specified in Input File (page 114). If not supplied, the current logscan input file is returned.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. This command added in version 3.7.1.
113
Input File
avmaint logscan input files must conform to the following format:
<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <logscanfilelist> <logscanfile pathname="/var/log/messages" index="0"> <logscanpattern order="0" severity="error" time="%b %d %T" pattern="^([a-zA-Z]{3} [0-9 ][0-9] ([0-9]{2}:){2}[0-9]{2}).*error" prefix="kernel error: "> <exclude pattern="<Code>.+<Type>.+<Severity>.+<Category>.+<HwSource& gt;.+<Summary>"/> </logscanpattern> <logscanpattern order="1" severity="info" time="%b %d %T" pattern="^([a-zA-Z]{3} [0-9 ][0-9] ([0-9]{2}:){2}[0-9]{2}).*kernel:" prefix="kernel info: "> <exclude pattern="<Code>.+<Type>.+<Severity>.+<Category>.+<HwSource& gt;.+<Summary>"/> </logscanpattern>
The following table describes each avmaint logscan input file XML attribute:
ATTRIBUTE DESCRIPTION
logscanfilelist
This element contains one or more log file descriptors (that is logscanfile sub-elements) that describe which hardware log files to scan, which entries in those log files are of interest and how to format and output entries of interest as Avamar event codes. IMPORTANT: The logscanfilelist is limited to a maximum of eight log file descriptors.
Full path and filename of a log file that will be scanned. This index attribute provides an internal identifier for saving log file state. Valid values are whole integers 0 thru 7. NOTE: /var/log/messages should be index 0 for compatibility with previous server versions.
logscanfile prefix
Optional text string that is prefixed to each Avamar event generated from a detected pattern in the scanned log file. If true or not present, this log file is scanned. If false, log file is not scanned.
logscanfile enabled
114
logscanfile order
Optional attribute that specifies processing order for each matching pattern in a log file descriptor. Valid values are unsigned integers. Patterns are tested in the order specified by the optional order attribute, or in some undefined order if no order attributes are provided. Once a pattern is matched, no further testing is performed. All patterns without the optional order attribute will be tested after any patterns that specify an order attribute.
logscanpattern exclude
Optional attribute that specifies a list of POSIX 1003.2 regular expressions that will be excluded (ignored). Format this entry of interest as one of the following Avamar event types: error info msg warning Error. Informational. Message. Warning.
logscanpattern severity
logscanpattern time
Optional time stamp before which server will ignore entries in scanned log files. This allows the server to avoid generating events for entries that occurred prior to the time the server was started or restarted. If not included, an Avamar event will be generated for all matching log file entries. The time attribute should provide a descriptor (see strptime(3)) to interpret the entry time. The time subexpression must be the first subexpression in the pattern.
logscanpattern pattern
A POSIX 1003.2 regular expression that will be evaluated to determine log file entries of interest. All pattern matches use the modern extended regular expressions and ignore case. The regular expression should include a subexpression that identifies the time of the entry and the time attribute should provide a descriptor to interpret the entry time. The time subexpression must be the first subexpression in the pattern. The time allows the server to avoid generating messages for events that occurred prior to the time the server was started or restarted. If no time information is provided, a message will be generated for all matching log file entries
115
Optional text string that is prefixed to each log file entry detected by this pattern. If true or not present, this pattern will be matched. If false, pattern is not matched.
116
avmaint lookup
The avmaint lookup command looks up the supplied HASH and returns information about the associated index and data stripes. IMPORTANT: The avmaint lookup command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint lookup commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint lookup HASH [H | P | U] --avamaronly GLOBAL-OPTIONS
Parameters
HASH HASH is a 40-character hex hash value.
Command Options
H | P | U One of the following hash types can also be supplied, which restricts the lookup operation to specific areas of the server: H P U Hash Filesystem (HFS). Persistent store hash. User account hash.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. This command added in version 3.5.
117
avmaint ls
The avmaint ls command all objects in a specified location and below.
Synopsis
avmaint ls LOCATION [--long] GLOBAL-OPTIONS
Parameters
LOCATION List all objects in this LOCATION and below.
Command Options
--long Displays details about objects.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
This command added in version 2.0.2. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
118
avmaint lscp
The avmaint lscp command lists available checkpoints.
Synopsis
avmaint lscp [--full] [--keepmin] GLOBAL-OPTIONS
Command Options
--full --keepmin Lists all checkpoints in the system (including invalid checkpoints). Enables the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. This option added in version 4.0.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
Output
Status is output in XML in the following format:
<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <checkpointlist nodecount="8"> <checkpoint tag="cp.20071029223027" isvalid="true" refcount="8" cpctime="1193697027" nodestotal="8" stripestotal="4461" hfsctime="1190066259" dirstotal="4" deletable="false"> <hfscheck starttime="1193697089" nodestarttime="1193697525" nodefinishedtime="1193701649" validcheck="true" errors="0" type="full" stripes-checking="4451" stripes-completed="4451"> <hfscheckerrors/> </hfscheck> <nodeidlist count="8"> <nodeidrange
119
Number of nodes in system checkpoint. Name of checkpoint. If TRUE, this is a usable checkpoint (that is, one that can be HFS checked and if it passes, can be used as a reliable system rollback point). Number of nodes participating in checkpoint. Time at which checkpoint was taken. Number of nodes in system. Total number of stripes stored in checkpoint. Time at which HFS was initialized. Number of directories on each node participating in checkpoint. These generally correspond to the number of disk partitions.
Indicates if checkpoint can be automatically deleted. Time at which hfscheck was initiated. Earliest time at which check proper started on a node. Latest time at which check completed on a node. If TRUE, this HFS check was successful; If FALSE, this HFS check failed. Number of detected errors. NOTE: This value does not reflect any automatic data stripe repair that took place after the HFS check completed. Refer to Automatic Stripe Repair (page 22) for additional information.
120
type
Check type (page 22). One of the following: full reduced partial All checks are performed on all stripes. Checkpoint was taken on N nodes but checked on N-1 nodes. All other kinds of checks.
stripes-checking
Total number of stripes being checked. NOTE: This is always less than the number of stripes contained in the checkpoint because management stripes are never checked and not counted.
stripes-completed
Total number of stripes checked. NOTE: This value simply records the number of stripes that checked and does not include any information on how they were checked.
hfscheckerrors nodeidlist
Reported errors. nodeidrange is a subrange of nodes included in this checkpoint. dc lseqno useqno Datacenter ID. Lowest logical node number covered by this range. Highest logical node number covered by this range.
121
avmaint nodelist
The avmaint nodelist command returns status of all nodes.
Synopsis
avmaint nodelist [--short] GLOBAL-OPTIONS
Command Options
--short Returns a very abbreviated (short) report that does not contain configuration, checkpoint, garbage collect or HFS check information.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
Output
Status codes are: Online Offline Ready Migrating Dead Node is functioning properly. Node has experienced a problem. Transitional state that might or might not be due to normal operation. Node is migrating stripes. Decommissioned.
The full server access mode is typically represented as three four-bit fields. For example: mhpu+mhpu+0000 The most significant bits show server privileges, the middle bits show root user privileges and the least significant bits show privileges for all other users.
122
avmaint nodelist COMMAND REFERENCE The individual bits in these fields convey the following information: m h p u Maintenance allowed. HFS is writable. Persistent store is writable. User accounting is writable.
Examples
Although the following example continues (wraps) to more than one line, all commands and options must be entered on a single command line (no line feeds or returns allowed). This command returns status of all Avamar server nodes: avmaint nodelist --id=admin@avamar --hfsaddr=avamar-1.example.com --xmlperline=1
123
avmaint perf
The avmaint perf status command returns operational status and performance monitoring statistics for the entire server or specified nodes. The avmaint perf reset command resets the performance monitoring statistics. TIP: The showperfhistory (page 376) program runs the avmaint perf status command and displays the average disk read performance rates in an easy-to-view format, sorted first by event sets, then by average read rate.
Synopsis
avmaint perf {reset NODE-ID --disknum=NUM --events=EVENT-BITS | status [NODE-ID]} GLOBAL-OPTIONS
Commands
One (and only one) of the following commands must be supplied on each command line: reset NODE-ID --disknum=NUM --events=EVENT-BITS Resets the performance monitoring statistics. Where NODE-ID specifies a logical node number, --disknum=NUM specifies a particular disk on that node and --events=EVENT-BITS is a valid numeric events bit value for this disk. TIP: In order to determine a valid --events=EVENT-BITS value, run showperfhistory (page 376) and use the value in the Event Bits column for the particular disk you want to reset. NODE-ID, --disknum=NUM and --events=EVENT-BITS are all required. status [NODE-ID] Outputs performance status in XML format. If one or more NODE-IDs are supplied, then data is returned for those nodes only; otherwise, data for all nodes is returned.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line. This command added in version 4.1.
124
Output
Output is XML in the following format:
<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE perfstatuslist> <perfstatuslist> <perfstatus node="0.F" create-time="1155949181" start-time="1155947404"> <stripekinds indx="4" data="8" comp="2" winx="1" wdat="1" wcmp="1" uinx="1" udat="2" lpar="6" mang="1" dlst="4"/> <checkpoint count="2" time="2" stripe-total="62"/> <garbagecollect count="90" time="304"/> <hfscheck count="1" time="35"/> <crunch count="0" time="0" nchunks="0" nbytes="0"/> <connections> <connection type="unknown" count="2" time="2"/> <connection type="avmaint" count="7" time="8"/> <connection type="accounting" count="2" time="2"/> <connection type="restore+accounting" count="1" time="6"/> </connections> <addhashdata total-bytes="82055208"> <chunktype name="atomic" is-dir="false" is-encrypted="false" has-hints="false"> <chunksizes block-length="1024"> <size count="15" start="0"/> <size count="11" start="1024"/> <size count="10" start="2048"/> <size count="14" start="3072"/> <size count="1" start="61440"/> </chunksizes> </chunktype> <chunktype name="recipe2" is-dir="false" is-encrypted="false" has-hints="true"> <chunksizes block-length="1024"> <size count="59" start="0"/> </chunksizes> <compsizes> <elem count="8" size="1"/> <elem count="21" size="3"/> <elem count="20" size="4"/> <elem count="8" size="6"/> <elem count="2" size="9"/> </compsizes> </chunktype> </addhashdata> <gethashdata total-bytes="5143844"> <chunktype name="atomic" is-dir="false" is-encrypted="false" has-hints="false"> <chunksizes block-length="1024"> <size count="15" start="0"/> <size count="11" start="1024"/> <size count="10" start="2048"/> <size count="14" start="3072"/> <size count="1" start="61440"/> </chunksizes> </chunktype> <chunktype name="recipe2" is-dir="false" is-encrypted="false" has-hints="true"> <chunksizes block-length="1024"> <size count="59" start="0"/> </chunksizes> <compsizes> <elem count="8" size="1"/> <elem count="21" size="3"/> <elem count="20" size="4"/> <elem count="8" size="6"/> <elem count="2" size="9"/>
125
126
avmaint perf COMMAND REFERENCE Each state attribute within the perfhistory element (for each disk) can have the following values:
VALUE MEANING
Disk is online and performing satisfactorily. Disk is suspended. Disk is in the process of being suspended. Disk is in the process of being put back online.
<perfhistory diskid="1" devsize="737840967680" devname="/dev/sdb1" bufsize="1048576" timeoutsecs="300" tests-used="2676" tests-skipped="158" tests-failed="34" state="online">
127
avmaint ping
The avmaint ping command returns stripe status.
Synopsis
avmaint ping [{--alwaysping | --details}] [--kind={comp | data | dlst | indx | lpar | mang | ppar | spar | udat | uinx | wcmp | wdat | winx}] [--state={ONLINE | OFFLINE | OFFLINE_MEDIA_ERROR | READY | MIGRATING | RESTARTING}] GLOBAL-OPTIONS
Command Options
--alwaysping | --details Supplying either --alwaysping or --details forces a ping of all stripes in order to obtain the most current status. Default value is false. Returns status for all stripes of a particular kind. Valid values are one of the following: comp data dlst indx lpar mang ppar spar udat uinx wcmp wdat winx Hash composite data stripe. Hash atomic data stripe. Delete stripe. Hash index stripe. Local parity stripe. Manage stripe. Parity/parity stripe. Safe parity stripe. User data stripe. User index stripe. Read/write composite data stripe. Read/write atomic data stripe. Read/write index stripe.
--kind={comp | data | dlst | indx | lpar | mang | ppar | spar | udat | uinx | wcmp | wdat | winx}
128
avmaint ping COMMAND REFERENCE --state={ONLINE | OFFLINE | OFFLINE_MEDIA_ERROR | READY | MIGRATING | RESTARTING} Returns status for all stripes in a particular state. Valid values are one of the following: ONLINE Only return status for stripes in an online state. Only return status for stripes in an offline state. Only return status for stripes that are offline due to hard disk (media) errors. Only return status for stripes in a ready state. Only return status for stripes that are currently migrating data to other stripes. Only return status for stripes that are restarting.
OFFLINE
OFFLINE_MEDIA_ERROR
READY
MIGRATING
RESTARTING
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Similar to the avmaint stats command (page 134). You should include --xml and the --xmlperline=NUM global option to specify XML output and to control the number of XML attributes per line, respectively.
129
avmaint rebuildstripe
The avmaint rebuildstripe command returns rebuilds offline stripes. IMPORTANT: The avmaint rebuildstripe command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint rebuildstripe commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint rebuildstripe {STRIPE-ID | --full} [--force] --avamaronly GLOBAL-OPTIONS
Parameters
One (and only one) of the following parameters must be supplied on each command line: STRIPE-ID --full Rebuilds only this specific stripe. Rebuilds all offline stripes.
Command Options
--force Causes all processes to stop immediately without sending any special information to the current client sessions. This will interrupt current users of the system.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option.
Examples
This command rebuilds a single offline stripe (0.0-B9): avmaint rebuildstripe 0.0-B9 This command rebuilds all offline stripes in the system: avmaint rebuildstripe --full
130
avamaint removebadhashes
The avamaint removebadhashes command removes a set of hashes that are tied to a specific verification token. Requires that avmaint findbadhashes (page 91) be run first in order to obtain the verification token and to place the server in read-migrate mode. IMPORTANT: The avmaint removebadhashes command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint removebadhashes commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint removebadhashes VERIFICATION --avamaronly GLOBAL-OPTIONS
Parameters
VERIFICATION VERIFICATION token, output from avmaint findbadhashes (page 91).
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
This command added in version 4.1.
131
avmaint rmcp
The avmaint rmcp command removes one or more checkpoints. IMPORTANT: The avmaint rmcp command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint rmcp commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
rmcp {CP-ID | --full} [--keepmin]
Parameters
One (and only one) of the following parameters must be supplied on each command line: CP-ID --full If a valid checkpoint ID (CP-ID) is supplied, only that checkpoint is deleted. Supplying --full invokes multiple checkpoint removal mode. First, all checkpoints are initially considered as eligible to be deleted and the checkpoints are chronologically sorted in order to determine which ones will be retained in the system. Next, a previously specified number of most-recent and validated (HFSchecked) checkpoints will be marked as ineligible for deletion. This behavior is controlled by the avmaint config cpmostrecent and avmaint config cphfschecked commands, respectively. Finally, any checkpoints still determined to be eligible for deletion are then deleted by the system. This option added in version 3.0.1. NOTE: CP-ID and --full are mutually exclusive.
Command Options
--keepmin Enables the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. This option added in version 4.0.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
132
avmaint sessions
The avmaint sessions command returns information on the current active client sessions; output displays in XML format.
Synopsis
avmaint sessions [--full] GLOBAL-OPTIONS
Command Options
--full Beginning with version 3.5.1, avmaint sessions does not show unknown sessions unless the --full option is supplied.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Beginning with version 2.0.2, this command no longer outputs plain-text; output is now strictly XML. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
133
avmaint stats
The avmaint stats command returns status of all stripes by sending a message to each stripe and receiving back some information about each stripe that responds.
Synopsis
avmaint stats [{--alwaysping | --details}] [--extended] [--kind={comp | data | dlst | indx | lpar | mang | ppar | spar | udat | uinx | wcmp | wdat | winx}] [--timing] [--xml] GLOBAL-OPTIONS
Command Options
--alwaysping | --details Supplying either --alwaysping or --details forces a ping of all stripes in order to obtain the most current status. Default value is false. Returns maximum (extended) information. Otherwise, only stripe ID, kind and status are returned for each stripe. Specifies a particular kind of stripe operate on. Valid values are one of the following: comp data dlst indx lpar mang ppar spar udat uinx wcmp wdat winx Hash composite data stripe. Hash atomic data stripe. Delete stripe. Hash index stripe. Local parity stripe. Manage stripe. Parity/parity stripe. Safe parity stripe. User data stripe. User index stripe. Read/write composite data stripe. Read/write atomic data stripe. Read/write index stripe.
--extended
--kind={comp | data | dlst | indx | lpar | mang | ppar | spar | udat | uinx | wcmp | wdat | winx}
134
avmaint stats COMMAND REFERENCE --timing --xml Shows communications timing information. This is not the default setting. Formats output in XML.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
Similar to the avmaint ping command (page 128). Support for non-XML output is likely to be discontinued in a future release. You should include --xml and the --xmlperline=NUM global option to specify XML output and to control the number of XML attributes per line, respectively.
Examples
This command lists contents of the client message store: avmaint stats --xml --xmlperline=1
135
avmaint stripels
The avmaint stripels command returns file data for specified list of stripes.
Synopsis
avmaint stripels STRIPE-LIST [--full | --safe] GLOBAL-OPTIONS
Parameters
STRIPE-LIST Specifies which stripes to operate on. If not supplied (default behavior), XML output is generated for the current active stripe and all checkpoints back to and including the first checkpoint where the stored file is different to the active stripe, known here as the first safe checkpoint. When a stripe is irremediably corrupted, this is useful in determining which checkpoint can safely be used in a rollback to avoid use of the corrupted stripe.
Command Options
--full Lists all checkpoints. --full and --safe are mutually exclusive. --safe Lists only the first safe checkpoint, if any. --safe and --full are mutually exclusive.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
This command added in version 3.6. The XML output provides the stripe's node, disk and kind, and for each displayed checkpoint, the filename, inode, file mode, link count, user ID, group ID, size in bytes, file block count and the standard file access, modified and create times. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
136
avmaint suspend
The avmaint suspend command temporarily halts client activities and denies new client requests.
Synopsis
avmaint suspend [--force] [--now] GLOBAL-OPTIONS
Command Options
--force Forcefully and immediately stops all processes without sending any special information to connected clients. This will interrupt current users of the system. If supplied, existing client sessions are gracefully terminated and no new client sessions are allowed. If not supplied, new client sessions are denied without suspending or terminating existing client sessions. Connected clients receive a SERVER_EXITING error immediately prior to the server closing the connection socket.
--now
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
137
avmaint test
The avmaint test command simulates various internal server hardware faults for testing purposes. IMPORTANT: The avmaint test command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint test commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint test {readerrors | writeerrors | ioerrors | networkerrors} --disknum=NUM [--latency=USECS] --nodenum=NODE --percent=NUM --avamaronly GLOBAL-OPTIONS
Parameters
One (and only one) of the following parameters must be supplied on each command line: readerrors writeerrors ioerrors networkerrors Simulate disk read errors. Simulate disk write errors. Simulate disk read and write errors. Simulate network errors.
Command Options
--disknum=DISK --latency=USECS --nodenum=NODE --percent=NUM Sets DISK number. Sets simulated network latency in microseconds (USECS). Sets NODE number. Sets percentage of operations for which to simulate errors.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option.
138
avmaint testintegrity
The avmaint testintegrity command tests integrity of a stripe and its parity group. Status for each of the stripes in the parity group is returned; returned status codes will be good, corrupt or unknown. IMPORTANT: The avmaint testintegrity command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint testintegrity commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint testintegrity STRIPE-ID [--repair] --avamaronly GLOBAL-OPTIONS
Parameters
STRIPE-ID Specifies which STRIPE-ID to test.
Command Options
--repair Asserts automatic rebuilding of a single corrupted stripe if other good stripes are available.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. This command added in version 3.0.1. IMPORTANT: This command cannot be used while backups are in progress.
139
avmaint timesync
The avmaint timesync command synchronizes server times with client. IMPORTANT: The avmaint timesync command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint timesync commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint timesync [--count=NUM] [--repair] [--timeout=SEC] --avamaronly GLOBAL-OPTIONS
Command Options
--count=NUM --repair --timeout=SEC Specifies number of iterations. Default value is 1000. Specifies that an attempt should me made to correct timing differences. Specifies number of seconds (SEC) before the operation is deemed to have timed out and will be terminated.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option. You should include the --xmlperline=NUM global option in order to control the number of XML attributes per line.
140
avmaint timing
The avmaint timing command times various server operations. IMPORTANT: The avmaint timing command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint timing commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint timing {addhashdata | gethashdata | ispresent | ping | refcheck} [--count=NUM] [--hashonly] [--maxsize=NUM] [--minsize=NUM] [--seed=NUM] [--timeout=SEC] [--trace] --avamaronly GLOBAL-OPTIONS
Parameters
One (and only one) of the following parameters must be supplied on each command line: addhashdata gethashdata ispresent ping refcheck Returns average elapsed time required to add a new hash to the hash filesystem. Returns average elapsed time required to read a hash from the hash filesystem. Returns average elapsed time required to verify whether or not a hash exists in the hash filesystem. Returns average elapsed time required to establish communication with a node. Returns average elapsed time required to verify integrity of the hash filesystem.
Command Options
--count=NUM --hashonly --maxsize=NUM --minsize=NUM --seed=NUM Specifies number of iterations. Default value is 1000. Specifies that data not be sent for timing messages. Specifies maximum chunk size. Default value is 8192. Specifies minimum chunk. Default value is 1024. Specifies random seed number (NUM) for timing.
141
avmaint timing COMMAND REFERENCE --timeout=SEC --trace Specifies number of seconds (SEC) before the operation is deemed to have timed out and will be terminated. Specifies that hash tracing should be performed.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
142
avmaint tracehash
The avmaint tracehash command traces the specified hash. IMPORTANT: The avmaint tracehash command is considered an advanced command that is normally intended for use by expert users only. You must include the --avamaronly option with all avmaint tracehash commands. If you are unsure about any aspect of this command, contact EMC Technical Support for additional information before using it.
Synopsis
avmaint tracehash HASH [--full] [--percent=NUM] [--remove] --avamaronly GLOBAL-OPTIONS
Parameters
HASH 40-character hex HASH value.
Command Options
--full --percent=NUM --remove Forces tracing of all hashes. Controls number (NUM) of hashes traced. Forces removal of the hash.
Global Options
Refer to Global Options (page 56) for a complete list of options that can be supplied with any avmaint command.
Notes
You must include the --avamaronly option.
143
avmgr
The avmgr program is a command-line interface to the Avamar account management system. IMPORTANT: avmgr is extremely powerful and can, if misused, completely and irrevocably corrupt an entire Avamar server. For this reason, avmgr is strictly reserved for use by EMC Technical Support only. Do not under any circumstances instruct non-EMC personnel to use avmgr without prior approval from EMC Technical Support.
Synopsis
avmgr {addu | chgc | chge | chgl | chgp | chgr | chgv | cpdb | delb | delu | getb | getc | getd | getl | getm | gets | getu | logn | movb | newd | newm | priv | resf | resp | resr} [--account={LOCATION | "ref{CID}"}] [--acnt={LOCATION | "ref{CID}"}] [--ci=INFO] [--cmd=COMMAND] [--date=DATE] [--expires=NUM] [--flagfile=FILE] [--format=FORMAT] [--help] [--hfsaddr=AVAMARSERVER] [--hfsport=PORT] [--id=USER@AUTH] [--ignoreconfig] [--label=LABEL] [--le=END] [--list-backups] [--loc="ref{CID}"] [--logfile=FILE] [--ls=START] [--mr=NUM] [--mvpath=LOCATION] [--p=PASSWORD] [--password=PASSWORD] [--path={LOCATION | "ref{CID}"}] [--pr=PASSWORD] [--pv=PRIVILEGE] [--quiet] [--retention-type={daily | monthly | none | weekly | yearly}] [--server=AVAMARSERVER] [--tryagain] [--u=NAME] [--ud=AUTH] [--usage] [--verbose] [--verbosity=LEVEL] [--version ]
Commands
One (and only one) of the following commands must be supplied on each command line: addu Creates a new Avamar user account and associates that user with a specific home location in the Avamar server. Use: --u=NAME to create the new user name. --p=PASSWORD (password) and --pr=PASSWORD (password retry) to set the new password. --pv=PRIVILEGE to set the new privilege level. --path=LOCATION to specify the home location. Users from an external authentication domain must supply --ud=DOMAIN to indicate the external authentication domain. --p=PASSWORD and --pr=PASSWORD are not supported when an external authentication domain is used.
144
avmgr COMMAND REFERENCE chgc Changes contact information. Use --ci=INFO to supply new contact information. chge Changes backup expiration time. Use --expires=NUM to supply new expiration time in days (number of days until backup expires). chgl Changes backup label. Use --date=DATE to identify a specific backup and --label=LABEL to supply the new label. chgp Changes user password. Use --u=NAME to specify which user to modify and --p=PASSWORD (password) and --pr=PASSWORD (password retry) to set the new password. chgr Changes backup retention type. Use --retention-type to designate backup as a daily, weekly, monthly or yearly. If --retention-type is not supplied or type is set to none, backup is not explicitly assigned an advanced retention type. This command added in version 4.0. chgv Changes user privilege level. Use --u=NAME to specify which user to modify and --pv=PRIVILEGE to set the new privilege level. cpdb Dumps the accounting database into a script file. Use --list-snapups to list of backups (roothashes, timestamps, size and percentnew) for each account. delb Deletes backup. Use --date=DATE or --label=LABEL to identify a specific backup. delu Deletes user. Use --u=NAME to specify which user to delete. Note also that this only removes a user from a single location in the system. If the user account was created with multiple (home) locations, then additional delu calls will be necessary. getb Returns (gets) backup list. Returns list of all backups sorted by date, with the latest backup listed first. Information includes label number, label name, number of bytes that were written (created), total number of bytes that comprise the backup, number of bytes found to be already present and not needing to be rewritten and the expiration value (number of days) where a zero indicates that the backup is to be kept indefinitely. Use --ls, --le and --mr to request a specific range or number of backups. Use --retention-type to request one or more specific retention types (daily, weekly, monthly or yearly).
145
avmgr COMMAND REFERENCE getc Returns (gets) contact information for the specified client. Use --account=LOCATION to specify which client to list. getd getl Returns (gets) a domain list. Returns (gets) a list of directories and clients. Use --account=LOCATION to specify which domain to retrieve. This does not perform a recursive search through sub-domains. getm Returns (gets) a client list. Similar to the getl command, but only returns clients. gets Returns (gets) statistics. Information includes accumulative statistics for all clients in the domain and monthly statistics, including the quantity of bytes backed up and the quantity of bytes already stored in the server. Use --account=LOCATION to specify which domain to retrieve. getu Returns (gets) a list of users. Returns the list of users at a specific account. This does not perform a recursive search through the accounts. logn movb Tests authorization. If the user exists in the system, privilege level and type of account are returned. Moves backup. Use --mvpath=LOCATION to specify the new LOCATION for the backup. newd Creates new domain. Use: --account=LOCATION to specify the name of the new domain. --ci=INFO to supply new contact information. TIP: Use all of the following options to create a new user account at the same time: --u=NAME to create the new user name. --p=PASSWORD (password) and --pr=PASSWORD (password retry) to set the new password. --pv=PRIVILEGE to set the new privilege level. newm Adds new client. Use: --account=LOCATION to specify new name and location. For example, /clients/MyClient. --ci=INFO to supply new contact information. TIP: Use all of the following options to create a new user account at the same time: --u=NAME to create the new user name --p=PASSWORD (password) and --pr=PASSWORD (password retry) to set the new password --pv=PRIVILEGE to set the new privilege level priv Returns a list of available privilege settings.
146
avmgr COMMAND REFERENCE resf resp Resolves internal representation of the path. Translates the (fixed) internal representation into the full text-based path name. Resolves path name to the block reference (blkref). Translates a textbased path name into the (fixed) internal representation. Hosts should save this internal representation in the (avamar.cfg) settings file with the blkref parameter rather than the account option (text form) to facilitate renaming of accounts. If the text version is saved and the host's path is changed, then automatic backups would fail. resr Resolves internal representation of the path. Translates the (fixed) internal representation into the text-based path name only to the parent level.
Options
--account={LOCATION | "ref{CID}"} Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. When used with the resf command, a CID can be passed in to obtain the location for that CID. Same as --acnt or --path. --acnt={LOCATION | "ref{CID}"} Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. When used with the resf command, a CID can be passed in to obtain the location for that CID. Same as --account or --path. --blkref=REFERENCE This is used primarily in configuration files. This representation is required to support renaming and moving of accounts. Contact information. INFO can contain any text up to 1K in size. It is intended to hold addresses, phone numbers and email addresses of people to contact for billing or configuration changes. --cmd=COMMAND Alternative mechanism for invoking the avmgr commands and options.
--ci=INFO
147
avmgr COMMAND REFERENCE --date=DATE Backup DATE. DATE is a numerical value, processed according to the following rules: If DATE is smaller than 1032, it is assumed to be a Unix time stamp Otherwise, it is assumed to be equal to the number of 100ns increments since January 1, 1601 --expires=NUM Used with chge command to specify number (NUM) of days from the creation date or timestamp that a backup should be retained in the system. The timestamp for a backup is assigned when it finishes. Any extension in retention time is added from the time of chge command execution. An expiration value of zero (0) means there is no end date for removing the backup. --flagfile=FILE Specifies a FILE containing a list of options and values that will be processed by this utility as if they were included on the command line. Default is /usr/local/avamar/etc/usersettings.cfg. Output format. Valid values are: browser html plain xml xmldoc Default format is plain. xmldoc includes an XML document header in addition to the requested information. --help --hfsaddr=AVAMARSERVER Shows full online help (commands, options and full descriptions) for this utility, then exits. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). This value is typically recorded in a configuration file. It can also be set by way of an environment variable or on the command line with each transaction. Default address is localhost (127.0.0.1). Same as --server=AVAMARSERVER.
--format=FORMAT
148
avmgr COMMAND REFERENCE --id=USER@AUTH Authenticate as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --label=LABEL --le=END Backup LABEL. Optional short text description. List end. Used with various query commands (getb, getc, getd, getl, getm, gets and getu) and --ls option to bound a range of results. END is context-sensitive. When used with getb, it accepts a date in either of the following formats: YYYY-MM-DD HH-MM-SS YYYYMMDDHHMMSS END can also be truncated to the desired resolution. When used with other query commands, it accepts an alpha or numeric value, which will constitute the end of the desired list. --list-backups Used with cpdb command to return a list of backups (roothashes, timestamps, size and percentnew) for each account. This option added in version 3.7. --loc="ref{CID}" --logfile=FILE --ls=START Used with newm command to specify a hierarchical location on the Avamar server for a specific CID. Log to this FILE. If no FILE value is supplied, default log file (avmgr.log) is used. List start. Used with various query commands (getb, getc, getd, getl, getm, gets and getu) and --le option to bound a range of results. START is context-sensitive. When used with getb, it accepts a date in either of the following formats: YYYY-MM-DD HH-MM-SS YYYYMMDDHHMMSS START can also be truncated to the desired resolution. When used with other query commands, it accepts an alpha or numeric value, which will constitute the beginning of the desired list. --mr=NUM Max record count. Used with various query commands (getb, getc, getd, getl, getm, gets and getu) to limit the size of the returned list. Used with movb command to specify the new LOCATION for the backup. The new path name is always relative to the user's current home location. Set PASSWORD.
--mvpath=LOCATION
--p=PASSWORD
149
avmgr COMMAND REFERENCE --password=PASSWORD --path={LOCATION | "ref{CID}"} PASSWORD for the --id=USER@AUTH account. Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. When used with the resf command, a CID can be passed in to obtain the location for that CID. Same as --account or --acnt. --pr=PASSWORD Confirm PASSWORD. Always used with --p=PASSWORD.
150
avmgr COMMAND REFERENCE --pv=PRIVILEGE User PRIVILEGE level. Valid values are: access Allows additions, changes or deletions of users (for example, avmgr). Allows addition of backups (for example, avtar -c). Allows creation of accounts and clients. Allows backups to be deleted. Controls whether the user is active or not. Allows all maintenance operations (for example, avmaint shutdown, avmaint config, avmaint rebuildstripe). Allows external users to view all data (ignore saved access control lists). Allows read maintenance data (for example, avmaint nodelist, avmaint cpstatus). Allows modify maintenance operations (for example, avmaint hfscheck, avmaint decommission). Allows accounts to be moved or renamed. Allows login without session ticket (that is, not controlled by MCS). Allows retrieval of backups, accounts, statistics, contact info lists (for example, avtar -x). Allows read of directory info only. This option added in version 3.5. Named collections (must be specified alone): admin Equivalent to create + read + backup + access + move + delete + enabled + noticketrequired. Equivalent to backup + enabled + noticketrequired. Equivalent to read + enabled + noticketrequired. 151
ignoreacls
maint
manage
readdir
backuponly readonly
avmgr COMMAND REFERENCE --quiet --retention-type={daily | monthly | none | weekly | yearly} Turns off both warnings and status messages. Used with chgr to assign an advanced retention type, which allows this backup to be managed using advanced retention policies and settings. Valid values are one of the following types: daily monthly none Explicitly designate this backup as a daily backup. Explicitly designate this backup as a monthly backup. Do not explicitly assign any retention type to this backup (that is, treat it as a normal ondemand backup). Explicitly designate this backup as a weekly backup. Explicitly designate this backup as a yearly backup.
weekly yearly
Default is none. If set to none or not supplied, backup is not explicitly assigned an advanced retention type and any existing retention type designation is cleared. Used with getb to retrieve backups of a particular type. This option added in version 4.0. --server=AVAMARSERVER AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). This value is typically recorded in a configuration file. It can also be set by way of an environment variable or on the command line with each transaction. Default address is localhost (127.0.0.1). Same as --hfsaddr=AVAMARSERVER. --tryagain --u=NAME If server is idle, try completing this operation later. This is the default setting. Avamar user NAME. Used when creating a new user or modifying a user's privilege level or password. External authentication (AUTH) system. Always used with --u=NAME. Shows abbreviated online help (commands and options only, no descriptions) for this utility, then exits.
--ud=AUTH --usage
152
avmgr COMMAND REFERENCE --verbose Sets verbosity level to verbose. Equivalent to --verbosity=verbose. --verbosity=LEVEL Sets verbosity LEVEL. Valid values are: debug error fatal info verbose warn Default value is info. --version Shows version, then exits.
Avamar-Only Commands
Avamar-only commands access advanced functionality that is normally reserved for use by EMC personnel only. You must include the --avamaronly option in order to use many of these advanced commands. IMPORTANT: Misuse of these advanced commands can cause loss of data. If you are unsure about any aspect of these commands, contact EMC Technical Support for additional information before using them. Adds backup. Disables block. Enables block. Moves/renames a block. Initializes accounting system.
153
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. You must include the --avamaronly option in order to use many of these advanced command-line options. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
Sets accounting system version. Enables Avamar-only commands (page 153) and options (page 154). Sets backup information. Sets directory containing Avamar binary files. Default value is /root. Sets local cache directory. Default value is /root/.avamardata. Same as --vardir=DIR. Sets client ID (obsolete). Default value is 0. Same as --clientid=CID. Sets client ID (obsolete). Default value is 0. Same as --cid=CID. Sets socket connect time-out. Default value is 60 seconds. Sets encryption method. Valid settings are: proprietary ssl:AES128-SHA ssl:AES256-SHA ssl No encryption. 128-bit Advanced Encryption Standard. 256-bit Advanced Encryption Standard. A special mode in which avmgr negotiates and uses the strongest encryption setting that the client can support.
Default setting is proprietary. --forceaddr --helpx Forces use of HFSADDR:HFSPORT. Prints help including extended flags.
154
avmgr COMMAND REFERENCE --hfsport=PORT --hostname=NAME --ignoreconfig --nagle --nat --noconfig --nouserinfo --pizza --sessionid=NUM --sid=NUM --singleconn --su=STATE --sysdir=PATH --syslog --threadstacksize=SIZE Sets Avamar server port number (HFSPORT). Default value is 27000. Sets name of the client machine (obsolete). Does not read any configurations files while parsing flags. Same as --noconfig. Turns on nagling. Sets whether the server is using NATing modules. Does not read any configurations files while parsing flags. Same as --ignoreconfig. Suppresses getting UID and GID mapping. Turns on debugging messages. Same as --debug. Sets session ID (obsolete). Same as --sid=NUM. Sets session ID (obsolete). Same as --sessionid=NUM. Makes a single connection. Sets update state. Directory containing Avamar server files. Sends log messages to the server. Default value is TRUE. Explicitly sets the default stack size. Default value is zero (0), which specifies that the operating system default stack size should be used. This option added in version 3.5. --vardir=DIR --wid=WID --workorderid=WID Sets local cache directory. Default value is /root/.avamardata. Same as --cachedir=DIR. Sets work order ID (obsolete). Same as --workorderid=WID. Sets work order ID (obsolete). Same as --wid=WID.
155
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --list-snapups IMPORTANT: Deprecated in version 4.1 in favor of --list-backups. Used with cpdb command to return a list of backups (roothashes, timestamps, size and percentnew) for each account. This option added in version 3.7.
Notes
Account Names Account names are case sensitive and can include only alphanumeric, underscore,
period and hyphen characters. Account names cannot be longer than 63 characters.
User Names User names are case sensitive and can include only alphanumeric, underscore, Passwords Passwords are case sensitive and can include only alphanumeric, underscore,
period and hyphen characters. User names cannot be longer than 31 characters. period and hyphen characters. Passwords must be at least six characters long but no longer than 31 characters and must contain at least one alpha and one numeral.
Examples
Although some of these examples continue (wrap) to more than one line, all commands and options must be entered on a single command line (no line feeds or returns allowed). This command creates a new directory in the users home location: avmgr newd --id=jdoe@avamar/clients/MyClient --password=ab4de6g --server=avamar-1 This command adds a new client and a user for that client: avmgr newm --id=jdoe@avamar/clients/MyClient --account=clients/bobsclient --password=ab3de6g --u=bob --p=PSWD --pr=PSWD --pv=backuponly This command adds a new user: avmgr addu --id=admin@site --account=clients/bobsclient --password=ab3de6g --server=avamar-1 --u=bob --p=PSWD --pr=PSWD --pv=readonly This command adds a new user from an external authentication domain: avmgr addu --id=admin@site --account=clients/bobsclient --password=ab4de6g --server=avamar-1 --u=bob --ud=ldap --pv=readonly This command changes a password: avmgr chgp --id=admin@site --account=clients/bobsclient --password=ab4de6g --server=avamar-1 --u=bob --p=NEWPSWD --pr=NEWPSWD
156
avmgr COMMAND REFERENCE This command returns a list of users: avmgr getu --id=admin@site --server=avamar-1 --password=ab4de6g --account=clients/bobsclient This command returns a list of directories: avmgr getl --id=admin@site --server=avamar-1 --password=ab4de6g --account=clients/bobsclient This command deletes a backup created on a specific date (--date) and Avamar server (--server): avmgr delb --id=admin@avamar/site/node16 --password=ab4de6g --server=avamar-1 --date=1019168580 This command returns a list of backups for the specified client: avmgr getb --id=admin@site --password=ab4de6g --server=avamar-1 --account=clients/bobsclient This command returns the CID for /clients/host1 in XML format: avmgr resp --path=/clients/host1 --format=xml
1 Request succeeded <account id="431de870.e4ac3f066775bd33898c2ac79bb78f6333488393" name="/clients/host1" />
157
avrpm_report
The avrpm_report program reports changes in the installed EMC RPM list. Default behavior is to output differences between master and current RPM lists.
Synopsis
avrpm_report [--current] [--debug] [--diff] [--dryrun] [--help] [--master] [--verbose]
Options
--current --debug --diff --dryrun --help --master --verbose Outputs current RPM list. Prints debugging information. Outputs difference between master and current RPM lists. Prints utility session information but does not actually perform the actions. Shows help, then exits. Outputs master RPM list. Provides maximum information.
158
avscc
The avscc program responds to plug-in browsing, backup and restore requests passed to it from the local avagent.bin (page 50) service and performs the actual work associated with these requests. It is typically located in Avamar Windows Client C:\Program Files\avs\bin directory, runs as a service on Avamar Windows Clients and is invoked by users from the Windows system tray. IMPORTANT: avscc is not intended to be run directly from the command line. The information provided in this topic is for reference purposes only.
Synopsis
avscc [--acaddr=IP-ADDR] [--acport=PORT] [--allowqueueworkorders] [--command={forceupdate | init | logfilelist | restore | snapup | status | stop | uninit | wo_restore | wo_snapup}] [--debug] [--dpnacl=USER@DOMAIN] [--dpnpath=PATH] [--encodings=NAME] [--flagfile=FILE] [--help] [--hfsaddr=IP-ADDR] [--informationals] [--interfacelanguage=NAME] [--logfile=FILE] [--log=FILE] [--mcsaddr=IP-ADDR] [--noinformationals] [--nostdout] [--nowarnings] [--quiet] [--server=IP-ADDR] [--snapset=NAME] [--snapupremindhours=NUM] [--usage] [--verbose | -v] [--version] [--wid=NAME]
Options
--acaddr=IP-ADDR Specifies IP address (IP-ADDR) that Windows client agent will use to directly communicate with the local avagent.bin service. Default value is 127.0.0.1 (localhost). --acport=PORT Specifies data PORT that the Windows client agent will use to directly communicate with the local avagent.bin service. Submits new workorders even when other workorders are currently running.
--allowqueueworkorders
159
avscc COMMAND REFERENCE --command={forceupdate | init | logfilelist | restore | snapup | status | stop | uninit | wo_restore | wo_snapup} Client agent command. One of the following: forceupdate init logfilelist restore snapup status stop uninit Activates this client with the specified Avamar server. Shows log file information. Performs a restore. Performs a backup. Shows status for all work orders related to this client. Stops the current operation. Removes activation with the specified Avamar server.
wo_restore wo_snapup --debug --dpnacl=USER@DOMAIN Prints utility session information but does not actually perform the actions. ACLs for the machine in the DPN accounting system. Where USER is the Avamar user name and DOMAIN is the domain name. --dpnpath=PATH --encodings=NAME --flagfile=FILE Specifies the directory (PATH) that identifies the machine. Specifies character encodings. Specifies a FILE containing a list of options and values that will be processed by avagent.bin as if they were included on the command line. Default is /usr/local/avamar/etc/usersettings.cfg. Shows help, then exits. NOTE: On the Windows client, avscc --help does not work. --hfsaddr=IP-ADDR Avamar IP address (IP-ADDR) or fully qualified host (as defined in DNS). Same as --server=IP-ADDR. --informationals Turns on all status messages.
--help
160
avscc COMMAND REFERENCE --interfacelanguage=NAME --log --logfile=FILE --mcsaddr=IP-ADDR --noinformationals --nostdout --nowarnings --quiet On Windows clients, NAME specifies the requested user interface language for the GUI. Logs to the FILE specified by --logfile=FILE. This is the default setting. Used with the --log option to specify the full path and filename of the alternative log file. Specifies the administrative server IP address (IP-ADDR). Turns off all status messages. Turns off output to STDOUT. Turns off warning messages Turns off both warnings and status messages. Equivalent to --noinformationals plus --warnings. Avamar IP address (IP-ADDR) or fully qualified host (as defined in DNS). Specifies a filename for the snapset. Specifies the number (NUM) of hours before a backup reminder message is displayed. The default is 24 hours. --usage Shows abbreviated online help (commands and options only, no descriptions) for this utility, then exits. Supplying either --verbose or -v turns on all messages (status and warnings). Shows version, then exits. Sets work order ID.
Notes
avscc.cfg (page 427) is a flag file that contains options that are passed to avscc when it is invoked. This flag file is located in the Avamar client var directory. avscc.cfg is typically located in C:\Program Files\avs\var on Windows clients and /usr/local/avamar/var on Unix clients.
161
avsetup_avamarbin
The avsetup_avamarbin utility creates the Avamar File System (AvFS) avamarbin directories, which contain the various platform-specific versions of the avacl utility (page 49).
Notes
This program added in version 3.7.
avsetup_ems
The avsetup_ems program configures an EMS. When run, it installs the Tomcat java application server in /usr/local/jakarta-tomcat-5.5.9/ and installs the Avamar Enterprise Manager web application in Tomcat. IMPORTANT: avsetup_ems is not intended to be run directly from the command line. The information provided in this topic is for reference purposes only.
Synopsis
avsetup_ems [--updatejavahome]
Options
--updatejavahome Updates JAVA_HOME environment variable to the version used by the EMS.
Notes
This program added in version 3.5.0. JAVA_HOME environment variable must be set to /usr/java/jre1.5.0_12.
162
avsetup_mccli
The avsetup_mccli utility configures the Avamar Administrator CLI. Refer to your Avamar Administrator Command Line Interface Programmers Guide and Reference Manual for additional information.
Synopsis
avsetup_mccli [--app_root PATH] [--help] [--java_home PATH] [--mcsaddr AVAMARSERVER] [--mcspasswd PASSWORD] [--mcsport PORT] [--mcsuserid USER-ID] [--use_defaults] [--user_root PATH]
Options
--app_root PATH --help --java_home PATH --mcsaddr AVAMARSERVER Sets application top level directory (that is $AVAMAR_ROOT). Default is /usr/local/avamar. Shows help, then exits. Set Java runtime directory. Default is /usr/java/jre1.5.0_12. Specifies which MCS to use to process Avamar Administrator CLI commands. This option added in version 4.0. --mcspasswd PASSWORD Specifies PASSWORD for the --mcsuserid account. This option added in version 4.0. --mcsport PORT Specifies MCS data PORT. This option added in version 4.0. --mcsuserid USER-ID Specifies Avamar administrative user account (USER-ID) that will be used to run Avamar Administrator CLI commands. This option added in version 4.0. --use_defaults If supplied, avsetup_mccli runs non-interactively and uses defaults for all settings. This option added in version 3.7. --user_root PATH Sets user root directory. Default is $HOME/.avamardata/var.
Files
avsetup_mccli updates the $AVAMAR_ROOT/lib/mcclimcs.xml default options file.
Notes
This program added in version 2.1.
163
avsetup_mcs
The avsetup_mcs program configures an MCS. It is typically run immediately after MCS software installation and prior to initializing the MCS by way of mcserver.sh --init. It has no effect if it is run after the MCS is started by way of mcserver.sh --start. The avsetup_mcs program performs the following actions: 1. Updates the JRE version and installation directory used by all MCS scripts. 2. Defines a specific Avamar server that all clients managed by this MCS will use for backups and restores. 3. Directs all internal MCS-to-Avamar server communications to a specified Avamar server hostname. 4. Creates all necessary MCS user accounts on the Avamar server and sets the MCUser account password. If avsetup_mcs is invoked more than once, it will only perform the first three actions. By default, avsetup_mcs runs interactively, prompting the user for any required values not already supplied on the command line. If all the parameters it requires are supplied on the command line or if the --noprompt option is supplied, it will run silently.
Synopsis
avsetup_mcs [--backuponlypass=PASSWORD] [--backuprestorepass=PASSWORD] [--error] [--help] [--hfsaddr=AVAMARSERVER] [--hfsport=PORT] [--java=DIR] [--lib] [--localdns=DNS] [--mcpass=PASSWORD] [--natextaddr=IP-ADDR] [--nocreateaccounts] [--nonat] [--noprompt] [--prefs] [--restoreonlypass=PASSWORD] [--rootpass=PASSWORD] [--run] [--runasanyuser] [--smtphost=SMTP-SERVER]
Options
--backuponlypass=PASSWORD --backuprestorepass=PASSWORD --error --help Specifies the password for the backuponly user account. Specifies the password for the backuprestore user account. Shows error messages if execution fails. Shows help, then exits.
164
avsetup_mcs COMMAND REFERENCE --hfsaddr=AVAMARSERVER Specifies AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). This is the Avamar server that all clients will be directed to for backups and restores. If a hostname is used, this name must be resolvable by all clients using the system. For this reason, it is usually a name resolvable by the corporate DNS or other name servers external to the Avamar server. NOTE: The MCS will attempt to resolve AVAMARSERVER hostname by way of external DNS and ping that host in order to verify communication with the rest of the system. However, it must be understood that this is no guarantee that the clients will also be able to do so. --hfsport=PORT --java=DIR --lib Specifies data port used for intra-server communication. Default value is 27000. Specifies JRE installation directory (DIR) used by MCS. IMPORTANT: EMC internal use only. If supplied, settings are written to the default (lib) version of the mcserver.xml preferences file that resides in /usr/local/avamar/lib/mcserver.xml. This is not the default setting. --localdns=AVAMARSERVER Specifies AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). If a hostname is used, this name must be resolvable by the MCS. In order for the MCS to monitor the Avamar server in the presence of network failure or loss of external name servers, the Avamar server internal IP address/fully qualified hostname should be used. The AVAMARSERVER IP address or fully qualified hostname has a .local.avamar.com domain suffix, which is resolvable entirely by the name server on the Avamar utility node. You can also set this parameter to contain the same value as --hfsaddr to make use of external name servers. The Avamar server internal fully qualified hostname is set as the default setting. --mcpass=PASSWORD Specifies MCUser user account PASSWORD.
165
avsetup_mcs COMMAND REFERENCE --natextaddr=IP-ADDR If Avamar server is deployed in a Network Address Translation (NAT) environment, this is the actual (non-translated) utility node or single node server IP address. If supplied, all NAT related user prompts are suppressed during the avsetup_mcs session. This option added in version 2.0. --nocreateaccounts --nonat If supplied, creation of the MCS user accounts on the Avamar server is bypassed. If supplied, avsetup_mcs does not prompt for the NAT external address and inserts a null value into the mcserver.xml external_nonat_addr preference. This option added in version 3.7. --noprompt --prefs If supplied, avsetup_mcs runs silently. IMPORTANT: EMC internal use only. If supplied, settings are written to the active (server_data) version of the mcserver.xml preferences file that resides in /usr/local/ avamar/var/mc/server_data/prefs/ mcserver.xml. This is the default setting. Supply --noprefs to turn this off. --restoreonlypass=PASSWORD --rootpass=PASSWORD Specifies the password for the restoreonly user account. Specifies Avamar server accounting system root user PASSWORD. MCS uses this password to log into the Avamar server to create all the necessary accounts it needs and for requesting monitoring data. IMPORTANT: EMC internal use only. If supplied, this utility can be run by any user. Specifies IP address or hostname (as defined in DNS) of the outgoing Simple Mail Transfer Protocol (SMTP) mail server (SMTP-SERVER) that will be used to send email home status messages.
Examples
This example sets the NAT address directly from the command line and runs silently, but does not affect any other settings: avsetup_mcs --extnataddr=x.x.x.x --noprompt
166
avsetup_mds
The avsetup_mds program enables the Avamar metadata search feature. It performs two actions: First, it creates a symbolic link from the Avamar File System (AvFS) mount point /mnt/axion to /var/www/html/axion. This symbolic link allows the Avamar File System (AvFS) to be accessed by way of Avamar Enterprise Manager so files returned by the search can be displayed as a hyperlinked list. Secondly, the default schedule for the starting metadata_cron (page 317) is inserted into the MCS database.
Synopsis
avsetup_mds [--accessnodehost=NAME] [--mdspass=PASSWORD] [--mdsdbport=PORT] [--testconfigured] [--testmdsrunning] [--updateaccessnode]
Options
--accessnodehost=NAME --mdsdbport=PORT --mdspass=PASSWORD Specifies access node hostname. Specifies metadata search database TCP port. Sets new mdsuser account PASSWORD. This option added in version 4.0. --testconfigured Tests whether or not the Avamar metadata search feature has been configured. If it has been configured, avsetup_mds exits with return code 0. If it has not been configured, avsetup_mds exits with a non-zero return code. --testmdsrunning Tests whether or not the Avamar metadata search feature is running. If it is running, avsetup_mds exits with return code 0. If it is not running, avsetup_mds exits with a nonzero return code. --updateaccessnode Updates access node hostname and port configuration. This option added in version 3.7.1.
Notes
This program added in version 3.5.1. The --testconfigured and --testrunning options are intended to be used by other scripts that can process the return codes to determine the status of metadata search
167
avsetup_snmp
The avsetup_snmp program configures a Net-SNMP agent so that it can communicate with an Avamar server by way of the Avamar Simple Network Management Protocol (SNMP) sub-agent. When avsetup_snmp is run, it examines an existing /etc/snmp/snmpd.conf file (if present) for configuration settings required by the Avamar sub-agent. If no v1/v2c SNMP communities are configured, then snmpconf --access_control is run to set up the required read/write and readonly SNMP communities. If system configuration settings are not present in /etc/snmp/snmpd.conf, then snmpconf --system_setup is run. Finally, snmpd is enabled.
Synopsis
avsetup_snmp [--access_control] [--config_dir=DIR] [--mibdir=DIR] [{--on | --off}] [--port=PORT] [--system_setup] [--verbose]
Options
--access_control --config_dir=DIR --mibdir=DIR --on | --off Forces running of snmpconf --access_control command during utility session. Specifies an alternate location for the snmpd.conf file. Specifies an alternate location for the Avamar MIB file. Supplying --on enables snmpd. This is the default setting. Supplying --off disables snmpd. --port=PORT --system_setup --verbose Specifies an alternate data PORT for SNMP communication. Default port is 161. Forces running of snmpconf --system_setup command during utility session. Provides maximum information.
Notes
IMPORTANT: Do not run avsetup_snmp from /etc/snmp. Doing so will cause any existing /etc/snmp/snmpd.conf file to be deleted. Configuration settings for snmpd are stored in /etc/snmp/snmpd.conf. avsetup_snmp will generate a basic snmpd.conf file. Another script (snmpconf ) is provided with net-snmp. The snmpconf script should only be run if you have solid knowledge of net-snmp and want to use Net-SNMP to monitor other SNMPenabled clients besides the Avamar server. Documentation for snmpconf is available by way of man pages provided with the net-snmp package or from www.net-snmp.org. AVAMAR 4.1 TECHNICAL ADDENDUM 168
avsetup_webstart
The avsetup_webstart program configures Avamar Administrator from Avamar Enterprise Manager.
169
avsetup_wrapper
The avsetup_wrapper performs selected subsystem initializations and updates.
Synopsis
avsetup_wrapper [--debug] [--help] [--verbose] {ems-init | ems-setup | help | lm-start | mccli | snmp-update | tomcat-update | website-update | webstart}
Commands
One (and only one) of the following commands must be supplied on each command line: ems-init ems-setup help lm-start mccli snmp-update tomcat-update website-update webstart Performs second part of EMS subsystem initialization. Performs first part of EMS subsystem initialization. Shows help, then exits. Same as supplying the --help option. Stops, then restarts Login Manager processes (lm). Initializes or updates mccli. Updates MCS SNMP configuration. Updates EMS Tomcat component. Updates web server configuration. Initializes or updates Avamar Administrator web interface.
Options
--debug --help --verbose Prints program session information but does not actually perform the actions. Shows help, then exits. Same as supplying the help command. Provides maximum information.
Notes
This program added in version 3.5.
170
avtar
The avtar program is a command-line backup and restore utility used to: Backup files and directories Delete an existing backup Extract and restore files or directories from a previous backup List the labels and dates of backups, or list the names of files and directories in a backup Validate a backup to ensure that data can be extracted
Synopsis
avtar {{--create | -c} | --delete | {--extract | --get | -x} | {--list | -t} | --backups | --validate} FILE1 [FILE2 ... ] DIR1 [DIR2 ... ] [--account=LOCATION] [{--aclrestore | --existing-dir-aclrestore}] [--after=DATE] [--backupsystem] [--before=DATE] [--count=NUM] [{--dereference | -h}] [--diff_sequencenumber=NUM] [{--directory=DIR | -C DIR}] [--exclude=PATTERN] [{--exclude-from=FILE | --exclude_from=FILE | -X FILE}] [{--existing-file-overwrite-option | --overwrite}={always | modified | never | newest | newname}] [--expires={NUM | TIMESTAMP}] [--flagfile=FILE] [--force] [--forcefs=FSTYPE] [--from-stdin] [--help] [--hfsaddr=AVAMARSERVER] [--id=USER@AUTH] [--ignorefs=FSTYPE] [--include=PATTERN] [{--include-from=FILE | --include_from=FILE}] [--informationals=N | --noinformationals] [--inflateofficexml={TRUE | FALSE}] [{--keep-old-files | -k}] [{--label=NAME | -f NAME}] [--labelnumber=ID] [--libavctl_path=PATH] [--logfile=FILE] [--no-recursion] [--nostdout] [--nowarnings] [{--one-file-system | -l}] [--open-file-restore-option={deferred | never | newest | newname}] [--password=PASSWORD] [--path=LOCATION] [--preservepaths] [--quiet] [--repaircache] [--reparse_limit=N] [--restorehidden={TRUE | FALSE}] [--restoreshortnames] [--restoresystem] [--restorewfp] [--retention-type={daily | monthly | none | weekly | yearly}] [--run-after-freeze=SCRIPT] [--run-after-freeze-clauses=STRING | --run_after_freeze_clauses=STRING] [--run-at-end=SCRIPT] [--run-at-start=SCRIPT] [--run-at-start-clauses=STRING | --run_at_start_clauses=STRING] [--sequencenumber=ID] [--server=AVAMARSERVER] [--showlog] [--skip-high-latency] [--streamformat=N] [--systemstatefile=FILE.bkf] [--target=PATH] [--testwfp] [{--to-stdout | -O}] [--totals] [--tryagain] [--usage] [{--verbose | -v | --verbose=2 | -vv}] [--version]
171
Commands
One (and only one) of the following commands must be supplied on each command line: --create | -c Creates a new backup. Either --create or -c can be used. Typically, you should include a list of files, directories or a path you want to backup. If you do not specify which files, directories or a path to backup, your entire local filesystem is backed up. Deletes an existing backup. Backups can only be deleted one at a time. By default, this command deletes the most recent backup. However, you can delete a specific backup by supplying --after=DATE, --before=DATE, --label=NAME, --retention-type or --sequencenumber=ID options. This command added in version 4.0. --extract | --get | -x Restores (extracts) files or directories from a previous backup. Either --extract or -x can be used. The extraction (restoration) defaults to the most recent backup, if no selection criteria is provided. By default, existing files are not overwritten during an extraction (restoration). If you want existing files to be overwritten, supply the --existing-file-overwrite-option. During an extraction, the path indicated in the syntax when the backups were created is the path to which the files are extracted. The options most often used with the --extract command include --label and --target=PATH, where PATH is the device and directory location to where the files or directories is extracted. --list | -t Lists the contents of a backup. Either --list or -t can be used. When used with the --verbose option, it returns file and directory permissions, size, creation date and time, as well as the file or directory name. By default, information from the most recent backup is returned. However, you can return information on other backups by filtering for a specific range of backup creation dates using --after=DATE or --before=DATE. Additionally, if you labeled the backup when you created it, you can also filter by that descriptive label using --label=NAME. --backups Lists all backup names and when they were created by a specific user account. Output is sorted by backup time (latest first) and returns size, creation date and time, backup label and the path that was backed up. The listing can be filtered for a specific range of creation dates by using --after=DATE or --before=DATE. --validate Validates the integrity of a previous backup by attempting a restore without actually restoring any data.
--delete
172
File/Directory List
If using the --create, --extract or --list commands, you can include a specific list of files or directories you want to backup or extract (restore). If you do not supply a list, your entire local filesystem is backed up or extracted (restored). FILE1 [FILE2 ... ] DIR1 [DIR2 ... ] One or more files or directories you want to backup (create), restore (extract) or list. Separate multiple entries with white space. NOTE: You can supply both a list of files and a list of directories on the same avtar command line. Common glob operators (wildcards) such as asterisk (*) and question mark (?) are allowed. Refer to Wildcards (page 39) for additional information. NOTE: When specifying files or directories, case sensitivity will vary according to the target computing platform you are backing up. Files or directories specified for Windows platforms are not casesensitive; files or directories specified for most other platforms are case-sensitive. Default behavior is to recursively process all subdirectories.
Options
--account=LOCATION Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. Same as --path=LOCATION. --aclrestore | --existing-dir-aclrestore Used with --extract to replace security settings (ACLs) on pre-existing directories.
173
avtar COMMAND REFERENCE --after=DATE Used with --delete, --extract, --list or --backups to only process backups created on or after this calendar DATE. Expected date/time format is: YYYY-MM-DD HH:MM:SS DATE can also be truncated to the desired resolution. --after interacts with both --before and --count in the following manner: 1. If a valid range of dates is specified by supplying both --before and --after, avtar will operate on all backups within that range. --count is superfluous. 2. If either --before or --after is supplied without the other, then avtar will operate on the number of backups specified by --count. If --count is not supplied, only one backup is processed. --backupsystem Used with --create to backup Windows system state along with the filesystem. This option is only applicable to Windows platforms. When specified, the following actions are performed during the backup: 1. avtar invokes NTBackup, which generates a SystemState.bkf file in the EMC \var directory. This file stores a snapshot of the Windows system state, as well as the Windows\Sysvol folder on servers using Active Directory. You can specify -systemstatefile=FILE.bkf to override the default SystemState.bkf file location. 2. avtar then pre-processes the SystemState.bkf file in order to maximize data de-duplication. 3. Finally, avtar performs the backup, which is assumed to comprise the entire filesystem, including the SystemState.bkf file. This option added in version 2.1.
174
avtar COMMAND REFERENCE --before=DATE Used with --delete, --extract, --list or --backups to only process backups created before this calendar DATE. Expected date/time format is: YYYY-MM-DD HH:MM:SS DATE can also be truncated to the desired resolution. --before interacts with both --after and --count in the following manner: 1. If a valid range of dates is specified by supplying both --before and --after, avtar will operate on all backups within that range. --count is superfluous. 2. If either --before or --after is supplied without the other, then avtar will operate on the number of backups specified by --count. If --count is not supplied, only one backup is processed. --count=NUM Used with --backups to only list this number (NUM) of backups. If this option is not supplied, default behavior is to list all backups. Refer to --after=DATE (page 174) and Refer to --before=DATE (page 175) for additional information about how --count interacts with those options. --dereference | -h --diff_sequencenumber=NUM Includes all files pointed to by symbolic links. Either --dereference or -h can be used. Specifies backup sequence number (NUM) for comparison. Used to compare the list of files in two backups. This option added in version 3.5.1. --directory=DIR | -C DIR Changes to this directory (DIR) prior to backup or restore. Either --directory=DIR or -C DIR can be used.
175
avtar COMMAND REFERENCE --exclude=PATTERN Excludes these files from a backup or restore. PATTERN is a single matching pattern. Common glob operators (wildcards) such as asterisk (*) and question mark (?) are allowed. Refer to Wildcards (page 39) for additional information. NOTE: When specifying exclusions, case sensitivity will vary according to the target computing platform you are backing up. Exclusions specified for Windows platforms are not case-sensitive; exclusions specified for most other platforms are case-sensitive. You can use multiple --exclude=PATTERN options on the same command line. However, each --exclude=PATTERN option can contain only one matching PATTERN. TIP: Use the --exclude-from=FILE option to pass in multiple file exclusions. --exclude-from=FILE | --exclude_from=FILE | -X FILE Specifies FILE containing a list of matching patterns used to specifically exclude files from a backup or restore. Common glob operators (wildcards) such as asterisk (*) and question mark (?) are allowed. Refer to Wildcards (page 39) for additional information. NOTE: When specifying exclusions, case sensitivity will vary according to the target computing platform you are backing up. Exclusions specified for Windows platforms are not case-sensitive; exclusions specified for most other platforms are case-sensitive. Either --exclude-from=FILE, --exclude_from=FILE or -X FILE can be used. Multiple exclude files can be used. However, each one must be preceded by a separate --exclude-from=FILE option.
176
avtar COMMAND REFERENCE {--existing-file-overwrite-option | --overwrite}={always | modified | never | newest | newname} Used with --extract to overwrite existing files during a restore (extraction). One of the following: always Replaces any existing client file if it exists in the backup. Does not restore the file if the date/time stamp of the backup file is the same as the local client file. Prevents client files from being overwritten. Only restores the file if the date/time stamp of the file in the backup is newer than the file on the local client. Restores the file but appends a version number to the restored file (for example, myfile.txt becomes myfile(1).txt when restored).
modified
never
newest
newname
Either --overwrite or --existing-file-overwriteoption can be used. This option added in version 1.2. --expires={NUM | TIMESTAMP} Specifies backup expiration date and time. If an integer (NUM) smaller than 32-bits is supplied, backup expires in that number of days. If a TIMESTAMP is supplied, backup expires on that calendar date and time. 32-bit integers are processed as a Unix TIMESTAMP; 64-bit integers are processed as a Windows TIMESTAMP.
177
avtar COMMAND REFERENCE --flagfile=FILE Specifies FILE containing a list of options and values that will be processed by this utility as if they were included on the command line. Default is /usr/local/avamar/etc/usersettings.cfg. Forces traversal of Network Filesystem (NFS) and Loopback Filesystem (LOFS) mounts. Forces traversal of this filesystem type (FSTYPE). Used with --create to backup a stdin stream. IMPORTANT: The --from-stdin option requires a very specific command-line syntax. Refer to the awv_install command-line example in the Avamar Server Software Installation Manual for additional information. --help --id=USER@AUTH Shows full online help (options and full descriptions) for this utility, then exits. Authenticate as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --ignorefs=FSTYPE --include=PATTERN Forces avtar to not traverse (that is, ignore) this filesystem type (FSTYPE). Includes these files in a backup or restore. NOTE: This option only includes files that would otherwise have been excluded with the --exclude=PATTERN option. PATTERN is a single matching pattern. Common glob operators (wildcards) such as asterisk (*) and question mark (?) are allowed. Refer to Wildcards (page 39) for additional information. NOTE: When specifying inclusions, case sensitivity will vary according to the target computing platform you are backing up. Inclusions specified for Windows platforms are not casesensitive; inclusions specified for most other platforms are case-sensitive. You can use multiple --include=PATTERN options on the same command line. However, each --include=PATTERN option can contain only one matching PATTERN. TIP: Use the --include-from=FILE option to pass in multiple file inclusions.
178
avtar COMMAND REFERENCE --include-from=FILE | --include_from=FILE Specifies a FILE containing a list of matching patterns used to specifically include files in a backup or restore. Common glob operators (wildcards) such as asterisk (*) and question mark (?) are allowed. Refer to Wildcards (page 39) for additional information. NOTE: When specifying inclusions, case sensitivity will vary according to the target computing platform you are backing up. Inclusions specified for Windows platforms are not case-sensitive; inclusions specified for most other platforms are case-sensitive. Either --include-from=FILE or --include_from=FILE can be used. Multiple include files can be used. However, each one must be preceded by a separate --include-from=FILE option. --informationals=N | --noinformationals Supplying --informationals=N sets information level. For example, --informationals=3. Supplying --noinformationals turns off all status messages. This option added in version 1.2. --inflateofficexml={TRUE | FALSE} Enables or disables special processing of Microsoft Office 2007 files. Default is TRUE. This option added in version 4.0. --label=NAME | -f NAME Used with --create to apply this short descriptive NAME to the new backup or delete operation. Used with other commands and options to search or filter behavior by this NAME. Either --label=NAME or -f NAME can be used. --libavctl_path=PATH --logfile=FILE Specifies full PATH to the libavctl shared library. Logs to this FILE. If no FILE value is supplied, default log file (avtar. log) is used.
179
avtar COMMAND REFERENCE --nostdout --nowarnings --one-file-system | -l Turns off output to stdout; output will still go to log file. Turns off warning messages. Used with --create to only backup files on the same local filesystem as the target directory. Either --one-file-system or -l can be used. --open-file-restore-option= {deferred | never | newest | newname} Used with --extract to specify how open client files are restored (extracted). Valid settings are: deferred Replaces open files the next time the client computer is rebooted. Does not ever overwrite an open file during a restore. Only restores an open file if the date/time stamp of the file in the backup is newer than the open file on the local client. Restores the open file but appends a version number to the restored file (for example, myfile.txt becomes myfile(1).txt when restored).
never newest
newname
This option added in version 1.2. --password=PASSWORD --preservepaths PASSWORD for the --id=USER@AUTH account. Used with --extract to restore the complete path of all files relative to the --target directory. Turns off both warnings and status messages. Equivalent to --noinformationals plus --nowarnings. Repairs local avtar cache files. This option added in version 2.0. --reparse_limit=N Specifies maximum number (N) of reparse point levels to traverse in an NTFS filesystem. Default value is 1. This option added in version 1.2.
--quiet
--repaircache
180
avtar COMMAND REFERENCE --restorehidden={TRUE | FALSE} Used with --extract to control whether or not hidden files (files with the hidden attribute set) should be restored. This option added in version 2.0. --restoreshortnames Used with --extract to restore short (8.3) Windows filenames. NOTE: This option is only applicable to Windows XP and Server 2003 and later version. If supplied with earlier versions, it has no effect. This option added in version 3.7. --restore-sparse-threshold=KB Used with --extract to specify number of consecutive zero bytes in KB that must exist in a particular file in order to create a sparse region in that file. This option does not apply when directing output to a stream. Setting this option to zero disables all sparse region checking. This option added in version 3.6. --restoresystem Used with --extract to restore system files. Refer to avw_install in the Avamar Server Software Installation Manual for additional information about which kinds of files avtar considers to be system files on various platforms and expected restore behavior. This option added in version 2.0. --restorewfp Used with --extract to restore files protected with Windows File Protection (WFP). This option added in version 3.5.
181
avtar COMMAND REFERENCE --retention-type={daily | monthly | none | weekly | yearly} Used with --create to assign an advanced retention type, which allows this backup to be managed using advanced retention policies and settings. Valid values are one of the following types: daily Explicitly designate this backup as a daily backup. Explicitly designate this backup as a monthly backup. Do not explicitly assign any retention type to this backup (that is, treat it as a normal ondemand backup). Explicitly designate this backup as a weekly backup. Explicitly designate this backup as a yearly backup.
monthly
none
weekly
yearly
Default is none. If set to none or not supplied, backup is not explicitly assigned an advanced retention type. This option added in version 4.0. --run-after-freeze=SCRIPT Runs SCRIPT after OTM freezes open files. IMPORTANT: This option can only be run against Windows clients. SCRIPT must be located in C:\Program Files\avs\etc\scripts. SCRIPT must have a .bat, .js or .vbs extension. This option added in version 2.0.2.
182
avtar COMMAND REFERENCE --run-after-freeze-clauses=STRING | --run_after_freeze_clauses=STRING Use these clauses (STRING) to invoke the --run-after-freeze=SCRIPT. Refer to Specifying Various Kinds of User Accounts on the Command Line (page 196) for additional information about allowable values for STRING. --run-at-end=SCRIPT Runs SCRIPT at end of avtar session. SCRIPT must be located in: /usr/local/avamar/etc/scripts on AIX and Linux /opt/AVMRclnt/etc/scripts on HP-UX and Solaris C:\Program Files\avs\etc\scripts on Windows On Windows clients, SCRIPT must have a .bat, .js or .vbs extension. There are no limitations on other client platforms. This option added in version 2.0.2. --run-at-start=SCRIPT Runs SCRIPT at beginning of avtar session. SCRIPT must be located in: /usr/local/avamar/etc/scripts on AIX and Linux /opt/AVMRclnt/etc/scripts on HP-UX and Solaris C:\Program Files\avs\etc\scripts on Windows On Windows clients, SCRIPT must have a .bat, .js or .vbs extension. There are no limitations on other client platforms. This option added in version 2.0.2. --run-at-start-clauses=STRING | --run_at_start_clauses=STRING Use these clauses (STRING) to invoke the --run-at-start=SCRIPT. Refer to Specifying Various Kinds of User Accounts on the Command Line (page 196) for additional information about allowable values for STRING.
183
avtar COMMAND REFERENCE --sequencenumber=ID Used with --delete, --extract, --list or --vaildate to specify the backup on which to operate. This option added in version 1.2 and supersedes --labelnumber=ID. --server=AVAMARSERVER AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Same as --hfsaddr=AVAMARSERVER. --showlog Shows session log for successful backups. This option added in version 2.0. --skip-high-latency Used with --create to enable special processing that will not include certain files and directories with high latency in the backup. Default is false. This option added in version 3.7.1. --streamformat=N --systemstatefile=FILE.bkf Used with --create and --extract to specify stream format. Default is none. Used with the --backupsystem option to override the default Windows system state backup file (FILE.bkf) location. This option added in version 2.1. --target=PATH Used with --extract to restore files to a different location (PATH) on the local filesystem. Lists all files protected with Windows File Protection (WFP). This option added in version 3.5. --to-stdout | -O Used with --extract to restore a single file to stdout. Either --to-stdout or -O can be used. --totals Used with --list to print total bytes listed. IMPORTANT: You must also supply either --verbose or -v to view --totals information. --tryagain If server is idle, try completing this operation later. This is the default setting. Shows abbreviated online help (options only, no descriptions) for this utility, then exits.
--testwfp
--usage
184
avtar COMMAND REFERENCE --verbose | -v --verbose=2 | -vv Supplying either --verbose or -v turns on all messages (status and warnings). Supplying either --verbose=2 or -vv turns on highly verbose messages including the avtar session ID and mount point information for each backup. Additional levels of verbosity can also be specified with -verbose=N, where N is the desired level of verbosity. --version Shows version, then exits.
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
Sets accounting system version. --acntver=2 is currently the only supported setting. Connects directly to all datacenters. Default value is TRUE. Connects directly to all nodes. Used with -c to specify a custom backup TAG. Used with -t and -x to list or extract the specified backup, respectively. This custom TAG is independent of the backup label. This option added in version 3.7.2.
--bigdirgroup=NUM
Controls how many files should be handled in each big directory group. By default, files are read in groups of 10,000 entries. A larger value consumes more memory but will require fewer passes of the directory.
185
avtar COMMAND REFERENCE --bigdirslotlimit=SLOTS Controls threshold at which avtar considers a directory to be a big directory (bigdir). This threshold is based on the byte size of the directory itself and therefore is an estimate of the number of files that might be contained within that directory. Default setting of 25,000 causes directories that contain at least 25,000 slots to go through the new nbackbigdirsmartdirwalk logic in order to break the directory into smaller pieces. When this threshold is exceeded and the new bigdir handling code is used, a new INFO message appears: INFO: Handling big directory: "DIR-NAME" Where DIR-NAME is the actual directory name being processed. Setting --bigdirslotlimit=-1 disables all of the nbackbigdir-smartdirwalk logic in order to facilitate debugging. --bindir=PATH Sets directory containing Avamar binary files. This directory is specified during Avamar client installation. Default value is /usr/local/avamar/bin. Sets number of bytes to read from one file at a time during a backup. Default value is 131072. Specifies number of buckets in the hash cache. Default value is 10. Sets local cache directory. This directory is specified during Avamar client installation. Default value is /root/.avamardata. Same as --vardir=DIR. Outputs more chunker information. Used with --compress to specify the numeric backup compression level. Larger values specify more compression. Adds this FILE to the backup as .system_info/comment. Enables backup compression. Sets composite block sticky byte magic number. Default value is 1152. Enables communication statistics every second. Sets socket connect time-out. Default value is 60 seconds.
--chunkstats --clevel=NUM
186
avtar COMMAND REFERENCE --consolidate Enables special storage mode that stores data associated with a composite contiguously on the same disk. This option added in version 4.0. --controlport=PORT Sets control port number. The control port is a socket connection that avtar uses to monitor stats requests. Sets numeric debug level. Default value is -1. Turns on debugging messages. Enables degenerate mode. If set, avtar will take a backup of the client filesystem but will not connect to the server. Deletes a single backup from the Avamar Server. The selection method to determine which backup to delete is the same as other avtar commands that operate on backups. Note that in order to reduce accidental deletions, the --force option must also supplied. This option added in version 3.7.2. --depth=NUM Sets level in chunktree at which to send tree to server. Default value is 2. Larger values might use excessive amounts of client memory. If supplied, file-cache holds the atime of each file and a new backup compares the current and previous atimes. If a mismatch is detected, then the file is re-scanned to detect ACL changes. This option added in version 3.7. --dirthreshold=MAGIC --dpnobj Sets directory block sticky byte magic number. Default value is 8192. Uses the dpnobj internal representation when listing the contents of a backup. This option added in version 3.5.1. --dumpformat Writes NDMP stream. Used by avndmp only. This option added in version 2.0. --dumplittle Writes NDMP stream with little-endian byte ordering. Used by avndmp only. This option added in version 2.0. --dumpnetapp_acls Writes NDMP stream with NetApp ACLs. Used by avndmp only. This option added in version 2.0.
--delete
--detect-acl-changes
187
avtar COMMAND REFERENCE --encrypt= {proprietary | ssl | ssl:AES128-SHA | ssl:AES256-SHA} Sets encryption method. Valid settings are: proprietary ssl:AES128-SHA ssl:AES256-SHA ssl No encryption. 128-bit Advanced Encryption Standard. 256-bit Advanced Encryption Standard. A special mode in which avtar negotiates and uses the strongest encryption setting that the client can support.
Default setting is proprietary. --filecachemax=MB --filestatmin=NUM --filestats=NUM --fixedatomsize=BYTES --fixedcomposite=COUNT --forceaddr --format=FORMAT --freezecachesize=MB Sets maximum file cache size in MB (negative value is fraction of RAM). Default value is -8. Sets minimum byte threshold for candidates in filestats. Default value is 1048576. Sets number of files recorded in .system_info/filestats. Default value is 500. Sets fixed size for atomic chunks (disables sticky). Sets number of hashes in a composite block (disables sticky). Forces use of HFSADDR:HFSPORT. Sets display FORMAT. Default value is xml. Used with Open Transaction Manager (OTM) on Windows platforms to specify the OFA cache size in MB (negative value is percent or hard drive used). Default value is -1. Used with OTM on Windows platforms to specify OFA time-out for freezing volumes. Default value is 120. Used with OTM on Windows platforms to specify OFA quiet wait for freezing volumes. Default value is 5. Sets maximum hash cache size in MB (negative value is fraction of RAM). Default value is -16. Prints help including extended flags. Sets Avamar server port number (HFSPORT). Default value is 27000.
--freezetimeout=SECS
--freezewait=SECS
188
avtar COMMAND REFERENCE --ignoreacls --ignoreconfig --ignoredefaultexcludes --ignorerwxmapping --ignorewinperm --incpartials --internal --level0 Solaris: ignores Solaris ACLS security descriptors. Does not read any configurations files while parsing flags. Same as --noconfig. Windows: ignores default excludes from registry. Windows: ignores mapping Windows security descriptors to Unix RWX. Windows: ignores restoring Windows security descriptors. Include partial backups in accounting system queries. Restores or displays internal (hidden) directory types (.system_info). Used with --create to not send is-present messages (assume all hashes not present on server). Creates a new backup view based on a list of files already on the server. This option added in version 3.5.1. --makeview_nosubsysteminfos={TRUE | FALSE} If set true, system information from individual backups is not merged when creating a backup view. Default is FALSE. This option added in version 3.6. --makeview_pluginmerge={TRUE | FALSE} If set true, individual backups are merged into a single backup view. Default is FALSE. This option added in version 3.6. --mapi --mapipassword=PASSWORD Uses MAPI protocol to communicate with an Exchange server. Used by avexchange only. PASSWORD for --mapiprofile=PROFILE. A password might not be required in all circumstances. Used by avexchange only. Communicates with Exchange server defined in this MAPI PROFILE. Used by avexchange only.
--makeview
--mapiprofile=PROFILE
189
avtar COMMAND REFERENCE --maxchunksize=BYTES --maxcompsize=MB Sets maximum size of an atomic chunk. Default value is 61440. Sets threshold at which a composite's elements will be chained together rather than invidually queued in the TODO queue for simultaneous execution. Default is 100 MB (that is, composites that span 100MB or more of data will be chained). This option added in version 3.7.1. --maxfile=MB --maxfilesize=MB If not zero, skips files larger than this size in Megabytes (MB). Imposes a maximum file size on files that will be backed-up. Default setting is 104857600MB, which sets the limit at 100TB. Therefore any file that reports a size greater than 100TB will be ignored by avtar. Sets maximum number of backup or restore chunks in the air. Default value is 80. Used with --parallel to set how much data each prefetch thread is allowed to read before blocking occurs. Default value is 2048KB (2MB). This option added in version 3.5. --memman=NAME --memmap --minchunksize=BYTES --mincompsize=BYTES --mindirsize=BYTES --nagle --nat --nearest --nocache --noconfig --nohiddendir --nouserinfo --numconns=NUM Sets memory debugging mode. Sets memory map source files. Sets minimum size of an atomic chunk. Default value is 1000. Sets minimum size of a composite chunk. Default value is 96. Sets minimum size of a directory chunk. Default value is 1024. Turns on nagling. Sets whether the server is using NATing modules. Restores from nearest module. Disables both file and hash caches. Does not read any configurations files while parsing flags. Same as --ignoreconfig. Does not write backup information on exit. Suppresses getting UID and GID mapping. Sets number of connections.
--maxpending=NUM --maxprefetch=SIZE
190
avtar COMMAND REFERENCE --numports=NUM --oktoclear --parallel Sets HFS port count. Enables clearing of local cache. Default value is TRUE. Asserts parallel processing of directories during backups. This option added in version 3.5. --permissions --ping --pluginport=PORT Preserves permissions. Default value is TRUE. Checks availability of server. Returns status=0 if reachable. Specifies avagent.bin communication PORT. NOTE: This value can only be set programmatically by avagent. --numthreads=NUM --randchunk=NUM --randseed=NUM --raw --refcheck --replicate Sets number of threads. Sends random chunks to server up to given number (NUM) of chunks. Sets random seed. Does not expand composites blocks. Checks hash references of comp and dir chunks. Intelligently duplicates contents of one Avamar server to another. Refer to replicate (page 341) for additional information. IMPORTANT: EMC internal use only. Implements replicate --reportonly mode. Refer to replicate (page 341) for additional information. This option added in version 4.1. --search-format Uses the format compatible with metadata search when listing the contents of a backup. This option added in version 3.5.1. --sessionticket=FILE --short-output | --short_output --singleconn --size=BYTES --statistics | --stats Specifies a FILE containing management consolesupplied session ticket. Generates shorter output with --backups. Makes a single connection. Sets size hint for client pipe command. Outputs chunk statistics.
--replicate-reportonly
191
avtar COMMAND REFERENCE --status=SEC --sysdir=DIR --syslog --tarfile Enables periodic status messages and sets number of seconds between messages. Directory (DIR) containing Avamar server files. Sends log messages to the server. Default value is TRUE. Used with --extract to restore files to a defined system tape device. IMPORTANT: The --tarfile option requires a very specific command-line syntax: --to-stdout is required, as well as piping the output to the Unix dd utility. For example: avtar -x clients/MyClient --to-stdout --avamaronly --tarfile | dd of=/dev/tape obs=20b --threadstacksize=SIZE Used with --parallel to explicitly set the default stack size. Default value is zero (0), which specifies that the operating system default stack size should be used. This option added in version 3.5. --threshold=NUM --throttle=MBPS Sets atomic block sticky byte magic number (NUM). Default value is 30000. Controls rate at which avtar sends data to the server. If --throttle=MBPS is supplied, avtar will pause as long as necessary after sending each packet in order to ensure that network usage does not exceed the specified maximum bandwidth; maximum bandwidth is specified in mega bits per second. For example, --throttle=5 uses half a 10Mbps connection, --throttle=0.772 restricts usage to one-half of a T1 link. --udp --uflagsdebug Uses UDP instead of TCP. Sets debug parsing. If set, prints each option as it is processed and each configuration file as it is parsed. Forces backup of unknown filesystems. Sets local cache directory. This directory is specified during Avamar client installation. Default value is /root/.avamardata. Same as --cachedir=DIR. Puts output in web-parsable form.
--unknown-okay --vardir=DIR
--web-format | --web_format
192
avtar COMMAND REFERENCE --windebug --workorder=FILE Windows: opens progress/status window. Specifies a FILE containing management consolesupplied work request.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --archives IMPORTANT: Deprecated in version 1.2 in favor of --backups. Retrieves list of all backups. --blkref=NAME IMPORTANT: Deprecated in version 1.2. Sets hash representation of path. --cfmerge IMPORTANT: Deprecated in version 3.7. Merges with another hash cache file and counts collisions. --cid=CID IMPORTANT: Deprecated in version 1.2. Sets client ID. Default value is 0. Same as --clientid=CID. --clientid=CID IMPORTANT: Deprecated in version 1.2. Sets client ID (obsolete). Default value is 0. Same as --cid=CID. --complete IMPORTANT: Deprecated in version 3.7. Marks the root hash as complete. --display IMPORTANT: Deprecated in version 1.2. Displays output of flags. --hfsaddr=AVAMARSERVER IMPORTANT: Deprecated in version 1.2. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Same as --server=AVAMARSERVER. --hostname=NAME IMPORTANT: Deprecated in version 1.2. Sets name of the client machine. --keep-old-files | -k IMPORTANT: Deprecated in version 1.2. Used with --extract to protect existing files from being overwritten during a restore. IMPORTANT: If this option is not supplied, default behavior is to overwrite all files, regardless of creation date, during a restore. This could potentially cause loss of data.
193
avtar COMMAND REFERENCE --labelnumber=ID IMPORTANT: Deprecated in version 1.2 in favor of --sequencenumber=ID. Used with --extract to search for or filter backups. ID is a unique identifier assigned to each backup by the server at the time the backup is taken. --listbackups IMPORTANT: Deprecated in version 1.2. Shows a list of all backups. --nodelonerror IMPORTANT: Deprecated in version 3.7. Suppresses deletion of files when errors occur. --nomail IMPORTANT: Deprecated in version 1.2. Inhibits mailing avscan output. --partial IMPORTANT: Deprecated in version 3.7. Adds this backup to the previous backup. --path=LOCATION IMPORTANT: Deprecated in version 1.2. Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. Same as --account=LOCATION. --pizza IMPORTANT: Deprecated in version 1.2. Turns on debugging messages. Same as --debug. --restorelength=BYTES IMPORTANT: Deprecated in version 4.0. When restoring (extracting) a single file, specifies length of restored file in BYTES. --restoreoffset=BYTES IMPORTANT: Deprecated in version 4.0. When restoring (extracting) a single file, specifies a file offset in BYTES. --run-at-endexit={TRUE | FALSE} IMPORTANT: Deprecated in version 4.0. If set true and script invoked by way of --run-at-end encounters an error, avtar will exit. Otherwise, avtar will continue to run. Default is TRUE. This option added in version 2.0.2. --run-after-freezeexit={TRUE | FALSE} IMPORTANT: Deprecated in version 4.0. If set true and script invoked by way of --runafter-freeze encounters an error, avtar will exit. Otherwise, avtar will continue to run. Default is TRUE. This option added in version 2.0.2.
194
avtar COMMAND REFERENCE --run-at-startexit={TRUE | FALSE} IMPORTANT: Deprecated in version 4.0. If set true and script invoked by way of --run-at-start encounters an error, avtar will exit. Otherwise, avtar will continue to run. Default is TRUE. This option added in version 2.0.2. --smartconn IMPORTANT: Deprecated in version 3.7. Uses best connection for each hash. --sessionid=NUM IMPORTANT: Deprecated in version 1.2. Sets session ID. Same as --sid=NUM. --sid=NUM IMPORTANT: Deprecated in version 1.2. Sets session ID. Same as --sessionid=NUM. --wid=WID IMPORTANT: Deprecated in version 1.2. Sets work order ID. Same as --workorderid=WID. --winstreams IMPORTANT: Deprecated in version 3.7. Windows: snaps up windows stream data files. --workorderid=WID IMPORTANT: Deprecated in version 1.2. Sets work order ID. Same as --wid=WID.
Notes
Run Locations The avtar program resides in and can be run from any of the following locations:
Single-node server or multi-node server utility node /usr/local/avamar/bin AIX and Linux client /usr/local/avamar/bin HP-UX and Solaris client /opt/AVMRclnt/bin Windows client C:\Program Files\avs\bin
User Name and If your Avamar user name and password are present in your .avamar file (page Password 426), then --id=USER@AUTH and --password=PASSWORD are not required on the
command line.
195
Examples
Although some of these examples continue (wrap) to more than one line, all options must be entered on a single command line (no line feeds or returns allowed). This command snaps up all files within the MyFiles and abcd directories and labels the backup jdoeFiles: avtar -c --label="jdoeFiles" MyFiles/ abcd/ --id=jdoe@avamar/clients/MyClient This command lists information about the three most recent backups created after the indicated date and time. Verbose (status and warning) messages are turned on: avtar --backups --verbose --count=3 --after="2000-09-01 04:30:15" --id=jdoe@avamar/clients/MyClient This command lists files and directories inside the backup labeled jdoeFiles created before the indicated date and time: avtar -t --verbose=2 /myfiles/rem --label="jdoeFiles" --before="2000-09-01 04:30:15" --id=jdoe@avamar/clients/MyClient Highly verbose (--verbose=2) messages are turned on. This command extracts into the old_newsletters directory all of the files found in the backup labeled newsletters that were created before the indicated date and time: avtar -xv --target="old_newsletters" --before="2001-03-24 15:00:00" --id=jdoe@avamar/clients/MyClient --label="newsletters" This command extracts into the old_newsletters directory those files found in the abcd and MyFiles directories in the backup labeled newsletters: avtar -xv --label="newsletters" --target="old_newsletters" abcd/ MyFiles/ --id=jdoe@avamar/clients/MyClient
196
avtar COMMAND REFERENCE This command shows how to specify a user account that has been created at the client level: avtar -c --id=jdoe@/DOMAIN/CLIENT --ap=PASSWORD Notice that the --account option is not required because the client was specified by the --id option.
197
Script Subprocessing
The following clauses provide additional script sub-processing capabilities such as setting additional environment variables, controlling argument parsing behavior and setting time-outs. These additional processing clauses are passed into the script subshell using the --run-after-freeze-clauses=STRING (page 183) or --run-at-start-clauses=STRING (page 183) command line options. desc=TEXT env=NAME=VALUE Descriptive TEXT for this clause. Default is noscript-text-description Defines environment variable NAME with this VALUE for the script environment. Default behavior is to only inherit environment variables already set in parent shell, plus the following four environment variables: $AVAMAR_SCRIPT_DESCRIPTION $AVAMAR_VARDIR $AVAMAR_BINDIR $AVAMAR_SYSDIR
$AVAMAR_SCRIPT_DESCRIPTION is the string stored in the desc=TEXT clause, or an empty string if desc=TEXT is not defined. $AVAMAR_VARDIR, $AVAMAR_BINDIR and $AVAMAR_SYSDIR contain values corresponding to avtar --vardir, --bindir, and --sysdir options, respectively. exit-on-error=BOOL skip-on-error=BOOL If BOOL is true, exit this process if the sub script exits with a non-zero exit code. Default is false. If BOOL is true, skip the next backup/restore component if an error is encountered. Default is false.
198
avtar COMMAND REFERENCE list=BOOL | stringlist-args=BOOL Specifies one of two possible parsing modes for --run-after-freeze-clauses=STRING and --run-at-start-clauses=STRING: If BOOL is false, split STRING into separate arguments for the script. Arguments must be separated by white space. If BOOL is true, use each argument STRING as a single argument to be passed into the script. Arguments can contain white space and must be separated by commas. False is the default setting, primarily to ensure backward compatibility. However, this restricts the arguments that can be passed into the subscript. This backward compatibility mode requires that only one argument string be present. For example: --run-at-end="scriptname arg1 arg2" However, using list parsing mode (with this clause set to true), the following script command line containing multiple arguments with spaces is allowed: --run-at-end="scriptname","arg1 with spaces","arg2 with spaces" use-cscript=BOOL If BOOL is true, script will be executed with Microsoft cscript.exe. Default is false, and unless use-cscript-raw clause is defined, the script will be executed with Microsoft cmd.exe /q /c. If BOOL is true, script will be executed with Microsoft cscript.exe /nologo. Default is false, and unless the use-cscript clause is specified true, the script will be executed with Microsoft cmd.exe /q /c. Specifies script time-out in seconds (SEC). Default is 3600 (1 hour). If the script does not return before the time-out has elasped, the script will be terminated and avtar will resume processing as if the script had returned an error. If BOOL is true, create a stdout pipe with script. Default is false If BOOL is true, create a stderr pipe with script. Default is false
use-cscript-raw=BOOL
timeout-seconds=SEC
create-stdout-pipe=BOOL create-stderr-pipe=BOOL
199
axionfs
The axionfs program is a system service that implements the Avamar File System (AvFS). IMPORTANT: The axionfs program is not intended to be run directly from the command line. In order to configure and control the Avamar virtual filesystem feature on your Avamar server, use dpnctl (page 239).
Synopsis
axionfs [--account=LOCATION] [--avamarmetadata={TRUE | FALSE}] [--avamarmetafile | --noavamarmetafile] [--backstats] [--cachestats] [--comstats] [--dateoption={0 | 1 | 2}] [--deltamode] [--flagfile=FILE] [--fuseoptions="OPTIONS-LIST"] [--help] [--id=USER@AUTH] [--largeprefetch=NUM-BUFS] [--latestonly] [--logfile=FILE] [--mountpoint=PATH] [--password=PASSWORD] [--quiet] [--server=AVAMARSERVER] [--singlethreaded] [--smallprefetch=NUM-FILES] [--usage] [{--verbose | -v | --verbose=2 | -vv}] [--version ]
Options
--account=LOCATION Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed. Used to enable (true) or disable (false) extended filesystem metadata, which can be used by the avacl utility (page 49) to restore filesystem metadata from tape backups. Default setting is true. This option added in version 3.6.
--avamarmetadata={TRUE | FALSE}
200
axionfs COMMAND REFERENCE --avamarmetafile | --noavamarmetafile Controls the format in which filesystem metadata is rendered in the Avamar File System (AvFS). Supplying --avamarmetafile stores filesystem metadata in .avamarmetafile files. This is the preferred behavior and is the default setting. Supplying --noavamarmetafile stores filesystem metadata in .avamarmetadata subdirectories. This is the pre-4.1 behavior, which has been deprecated due to slow performance. This option added in version 4.1. --backstats Causes various performance and network transfer statistics to be printed on axionfs unmount. Displays objcache stats every second. This option added in version 3.6. --comstats Enables communication statistics every second. This option added in version 3.6. --dateoption={0 | 1 | 2} Renders by-date directory contents in one of the following date/time formats: 0 Renders by-date directory contents in YYYYMMDD_HHMMSS-UTC format. This is the default setting. Renders by-date directory contents in YYYYMMDD_HHMMSS-TZ format where the time is the local time and TZ is the local time zone (for example, PDT) Renders by-date directory contents in YYYYMMDD_HHMMSS format where the time is in the local time and the implicit time zone is the local time zone.
--cachestats
For all these date/time formats: YYYY is a fourdigit calendar year, MM is a two-digit calendar month, DD is a two-digit calendar date and HHMMSS is a six-digit time stamp. --deltamode Causes the contents of backups after the oldest for each account to contain only the changed or new files and directories; essentially an incremental update view. Specifies FILE containing a list of options and values that will be processed by this utility as if they were included on the command line. Default is /usr/local/avamar/etc/usersettings.cfg.
--flagfile=FILE
201
axionfs COMMAND REFERENCE --fuseoptions="OPTIONS-LIST" OPTIONS-LIST is a space-delimitted list, comprising of any of the following options: -d -f -s Debug mode (implies -f ). Remain in the foreground. Single-threaded in the fuse layer. NOTE: The top-level axionfs --singlethreaded option is somewhat redundant, but operates at a different level. -o Comma-separated list comprising any of the following filesystem options: allow_other Allow system users other than the user mounting the filesystem to have access. If unlink() is called while a file is open, remove that file immediately. Issue larger than 4Kb reads (only valid on 2.4.x kernels). Accept the inode numbers that axionfs code assigns in getdir or stat.
hard_remove
large_read
use_ino
IMPORTANT: The -o options list must be supplied after -d, -f or -s options. For example: --fuseoptions="-s -f -o allow_other,use_ino" --help --id=USER@AUTH Shows full online help (options and full descriptions) for this utility, then exits. Authenticate as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar.
202
axionfs COMMAND REFERENCE --largeprefetch=NUM-BUFS This controls the number of 256KB buffers (NUM-BUFS) that are automatically pre-fetched when reading large files. Default setting is 8, yielding up to 8*256=2MB of read-ahead. Setting this option to 0 disables the large file read-ahead cache. This option added in version 3.6. --latestonly Presents only the latest backup for the user whose credentials are being used under the mount point, rather than creating the full hierarchy of domains, clients and backups. Specifies path to the (empty) directory that will be used as the mount point for the filesystem. PASSWORD for the --id=USER@AUTH account. Suppresses logging messages. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Wraps all axionfs callbacks with a mutex. This is not the default setting. Default setting changed to false (--nosinglethreaded) in version 3.6. --smallprefetch=NUM-FILES This controls the number of small (<32KB) files (NUM-FILES) that are immediately pre-fetched when a directory is read. Default setting is 1024 files. Setting this option to 0 disables the small file pre-fetch cache. TIP: When AvFS is used predominantly for interactive use, rather than as a file export mechanism, use --smallprefetch=0 to improve interactive response time. This option added in version 3.6. --usage --verbose | -v --verbose=2 | -vv Shows abbreviated online help (options only, no descriptions) for this utility, then exits. Supplying either --verbose or -v turns on all messages (status and warnings). Supplying either --verbose=2 or -vv turns on highly verbose messages including the avtar session ID and mount point information for each backup. Additional levels of verbosity can also be specified with --verbose=N, where N is the desired level of verbosity. --version Shows version, then exits.
203
Notes
This program added in version 3.5. --avamarmetadata and --avamarmetafile options are typically defined in the /usr/local/avamar/var/axionfs.cmd configuration file.
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them. Use dummy implementation of getattr callback for testing. Use dummy implementation of getdir callback for testing. Use dummy implementation of read callback for testing.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged: --hextime IMPORTANT: Beginning with version 3.7, this option has been deprecated for most situations in favor of --dateoption. However, it is retained for backward compatibility with EMC Centera filesystems. If supplied, renders by-date directory contents in hex time format.
204
backup_upgrade_files
The backup_upgrade_files program is a utility that backs up critical Avamar configuration files so that a site-specific customizations can be saved during system upgrades.
Synopsis
backup_upgrade_files [--diffdir=TMP-DIR] [--dir=PATH] [--help] [--verbose]
Options
--diffdir=TMP-DIR Compares currently active Avamar configuration files with corresponding files in a temporary directory (TMPDIR) and reports whether or not there are any differences. NOTE: Specific differences within files are not reported, just whether or not the files differ. --dir=PATH --help --verbose Specifies the logfile PATH. Shows help, then exits. Provides full information.
Notes
backup_upgrade_files backs up the following files: /data01/home/admin/.avamar /etc/hosts /etc/ldap.conf* /etc/nsswitch.conf /etc/ntp.conf /etc/pam.d/lm* /etc/pam.d/login /etc/pam.d/system-auth /etc/resolv.conf /etc/sysconfig/network /etc/sysconfig/network-scripts/ifcfg-eth0 /etc/yp.conf /usr/local/avamar/bin/cp_cron /usr/local/avamar/bin/evening_cron_run /usr/local/avamar/bin/gc_cron /usr/local/avamar/bin/hfscheck_cron /usr/local/avamar/bin/morning_cron_run /usr/local/avamar/etc/domains.cfg /usr/local/avamar/etc/repl_cron.cfg /usr/local/avamar/etc/usersettings.cfg AVAMAR 4.1 TECHNICAL ADDENDUM 205
Examples
The following command compares currently active configuration files with corresponding files in /tmp/backup_files and reports verbosely whether or not there are any differences: backup_upgrade_files --diffdir=/tmp/backup_files --verbose
206
btfix
The btfix program changes code offsets from back traces stored to log files (usually from avtar or gsan processes) to human readable output with method, source file and line number. It is primarily used to assist EMC engineering with debugging.
Synopsis
btfix -PROGRAMFILE LOGFILE
Example
The output is to stdout so you will want to redirect it to a file. For example if you want to btfix a gsan.log file: btfix -gsan gsan.log > gsan.btfix.log
207
capacity.sh
The capacity.sh shell script generates a capacity utilization report listing the amount of Avamar server data that is added and freed each day.
Synopsis
capacity.sh [--client=CLIENT-NAME] [--days=NUM] [--domain=DOMAIN-NAME]
Options
--client=CLIENT-NAME Limits scope of report to this CLIENT-NAME only. In this context, report output changes to show dataset used for each backup. Limits scope of report to only include the specified number (NUM) of days. Default is 30 days. Limits scope of report to this DOMAIN-NAME only.
--days=NUM --domain=DOMAIN-NAME
Notes
This program added in version 4.1.
Example
The following command lists capacity utilization for the past 10 days: capacity.sh --days=10
Date 2008-10-06 2008-10-07 2008-10-08 2008-10-09 2008-10-10 2008-10-11 2008-10-12 2008-10-13 2008-10-14 2008-10-15 Average New Data #BU 1364 mb 7 2827 mb 7 60186 mb 7 60187 mb 7 60187 mb 7 60196 mb 7 60187 mb 7 60187 mb 7 60187 mb 7 60189 mb 7 48570 mb Removed #GC 0 mb 1 0 mb 1 0 mb 1 0 mb 1 0 mb 1 0 mb 1 0 mb 1 0 mb 1 -46801 mb 2 -99541 mb 1 -14634 mb Net Change 1364 mb 2827 mb 60186 mb 60187 mb 60187 mb 60196 mb 60187 mb 60187 mb 13385 mb -39351 mb 136.4 mb ---------- ---------- ----- ---------- ----- ----------
Top 3 High Change Clients: -------------------------awk: cmd. line:2: warning: escape sequence `\%' treated as plain `%' Total for all clients 1062500 mb 100.0% MyServer-1 MyServer-2 MyServer-3 159818 mb 154664 mb 154650 mb 15.0% 14.6% 14.6%
208
capacity.sh COMMAND REFERENCE The New Data column lists the amount of data (in megabytes) that has been added to the Avamar server on a daily basis. The #BU column shows the number of backups or replications that occurred each day. The Removed column lists the amount of data (in megabytes) that garbage collection recovered. The #GC column shows the number times garbage collection ran each day. The Net Change column lists the amount of data (in megabytes) that were added or freed each day. Finally, a summary at the end of the report lists the three highest change rate clients for the time period of this report.
209
change-passwords
The change-passwords program is an interactive utility that assists system administrators with changing various operating user account passwords and Avamar server user account passwords, as well as generating and propagating new OpenSSH keys. The change-passwords program will interactively prompt and guide you through any or all of the following operations: Changing operating system login passwords for the admin, dpn and root accounts Creating new admin and dpnid OpenSSH keys Changing internal Avamar server passwords for the root and MCUser accounts
Synopsis
change-passwords [--debug] [--help] [--verbose]
Options
--debug --help --verbose Enables debug mode, which prints messages programmers can use to debug program execution. Shows help, then exits. Provides maximum information.
Notes
The change-passwords program should be run as user dpn because the dpn user account inherently retains old SSH keys (for example, the original factory SSH dpnid key), which might be required to change keys when new nodes are added to multi-node servers. The change-passwords program verifies whether or not you are running as dpn and issues a warning if you are not. However, it is possible to run changepasswords as user admin, provided that the correct keys are available. It is not necessary to pre-load any SSH keys before running change-passwords. The change-passwords program disallows quote marks and other shell metacharacters in passwords. The change-passwords program presently does not remind users to update the repl_cron.cfg file on replication sources when passwords are changed on a replication target. The change-passwords program fails if the current working directory is not readable. The work-around is to use su - dpn instead of su dpn if one is starting out as the root user. The change-passwords program will fail when OpenSSH refuses to proceed because a host key has changed, causing OpenSSH to suspect a man-in-themiddle attack. The remedy is to update or to remove ~dpn/.ssh/known_hosts when a node is changed out in place (the node's hostname and IP address are the same as before, but the node has a new host key, either because the node is a new replacement or (less likely) because someone re-initialized the host key). AVAMAR 4.1 TECHNICAL ADDENDUM 210
Troubleshooting Information
If change-passwords fails, examine /usr/local/avamar/var/change-passwords.log for detailed information regarding any issues or problems reported in the final change-passwords interactive messages. For example, on some older nodes, ownership of /usr/local/avamar/etc/ might be root:root. This will cause change-passwords to exit with an error because it cannot create a backup of the usersettings.cfg file. The reason change-passwords exited would only be apparent by examining the change-passwords.log file.
211
change_nodetype
The change_nodetype program configures a new node to function as a specific node type in an Avamar server. It resides in the /usr/local/avamar/src directory and should be run from that location.
212
check.dpn
The check.dpn program performs various kinds of system integrity checks. It can be run directly and is also called by start.dpn (page 382). The actual checks are performed by executing run instruction files located in these directories: check.dpn.d/cron/5min (--checks=cron/5min) check.dpn.d/cron/10min (--checks=cron/10min) check.dpn.d/inventory (--inventory) check.dpn.d/monitor (--monitor) check.dpn.d/preinstall (--preinstall) check.dpn.d/postinstall (--postinstall) These directories do not contain the actual check scripts; these directories contain run instruction files. The actual scripts reside in /usr/local/avamar/bin/check.dpn.d/init.d.
Synopsis
check.dpn [--checks=DIR,...] [--debug] [--exclude=CHECK,...] [--inventory] [--mail] [--monitor] [--postinstall] [--preinstall] [--quiet] [--results=DIR] [--select=CHECK,...] [--verbose]
Options
--checks=DIR,... Executes run instruction files located in this directory (DIR). Use a comma-separated list to define multiple directories. Enables debug mode, which prints messages programmers can use to debug program execution. Explicitly excludes this CHECK from being run. Use a comma-separated list to exclude multiple check scripts. Performs inventory checks by running scripts based on instructions in check.dpn.d/inventory. Inventory checks return information about specific hardware and software installed in the system. --mail Emails information to the utility node admin user account. Exact behavior is highly dependent on the individual checks being run. --monitor Performs monitoring checks by running scripts based on instructions in check.dpn.d/monitor. Monitoring checks return the status of various system components.
--debug --exclude=CHECK,...
--inventory
213
check.dpn COMMAND REFERENCE --postinstall Performs post-installation checks by running scripts based on instructions in check.dpn.d/postinstall. Post-installation checks are generally performed after installing or upgrading Avamar software to verify integrity of a deployed Avamar server. --preinstall Performs pre-installation checks by running scripts based on instructions in check.dpn.d/preinstall. Pre-installation checks are generally performed prior to installing or upgrading Avamar software. --quiet --results=DIR --select=CHECK,... Turns off all non-essential messages. Specifies directory (DIR) for the results file. Explicitly executes on those checks (CHECK) specified by this option (excludes all other checks). Use a comma-separated list to define multiple checks. The --select option only works for checks that have corresponding R* (run instruction) files. Without the R* file, check.dpn will not run the check, even if a check by that name exists. This is because the R* file contains vital information about how to run the test. --verbose Provides maximum information.
Notes
IMPORTANT: You must load either the dpnid or admin OpenSSH key before running this utility. This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
214
check.dpn COMMAND REFERENCE 4. Copy check.dpn.d/preinstall/Template to create a new run instruction filenamed R{nnn}newcheck. Where {nnn} is a sequence of digits (for example, 105). This helps to control the script's placement in the lexicographic run order. 5. Edit the run instructions file (for example, R105newcheck) so as to indicate how and on what types of nodes to run your new script. 6. Place your new run instructions file into check.dpn.d/{mode}. Where {mode} is one of the subdirectories preinstall/, postinstall/, monitor/, inventory/. You can instead place the file into a user-defined directory that can be selected by way of the --checks=DIR option (for example, check.dpn.d/cron).
Examples
This command runs all preinstall checks: check.dpn --preinstall This command runs only checkntp in preinstall mode: check.dpn --preinstall --select=checkntp This command runs status checks defined in the cron/5min subdirectory: check.dpn --checks=cron/5min
215
check.mcs
The check.mcs program performs various integrity checks on the MCS. It can be run directly and is called by mcserver.sh (page 309) when the --check or --init options are supplied. IMPORTANT: The check.mcs program must be run from the utility node in multi-node Avamar systems.
Synopsis
check.mcs {--poststart | --preavsetup | --preinit | --prestart | --testserver} [--verbose]
Commands
One (and only one) of the following commands must be supplied on each command line: --poststart --preavsetup --preinit --prestart --testserver Tests a running MCS. Tests utility node setup and software installs. Tests readiness to initialize the MCS. Tests readiness to start the MCS. Tests Avamar server response time.
Options
--verbose Provides maximum information.
216
checklib.pm
The checklib.pm file is a Perl module used by mcsmon_run (page 315), monmcs (page 323) and pingmcs (page 327).
clean.dpn
The clean.dpn program completely removes all customer data from all /data??/cur.* or /data??/hfscheck.* directories. IMPORTANT: The clean.dpn --cur command will completely remove all customer data from the system. Do not run this command unless instructed to do so by EMC Technical Support.
Synopsis
clean.dpn [--checkpoints={all | CP-ID,CP-ID,...}] [--cur] [--hfscheck] [--nodes=NODE-LIST] [--norollback] [--verbose]
Options
--checkpoints={all | CP-ID,CP-ID,...} Completely removes (cleans) one or more specified checkpoints (CP-IDs) or all checkpoints. This option added in version 4.1. --cur --hfscheck --nodes=NODE-LIST Completely removes (cleans) all customer data from all /data??/cur.* directories. Completely removes (cleans) all customer data from all /data??/hfscheck.* directories. Specifies which node(s) to clean. If not supplied, all nodes are cleaned. Requires physical MODULE.NODE designations as defined in probe.out (page 477). Refer to Physical vs. Logical Node Numbers (page 478) for additional information. --norollback --verbose Does not remove (clean) certain temporary files that might remain following a server rollback. Provides maximum information.
217
copy-ata-drive
The copy-ata-drive program makes a copy of an ATA drive containing Linux ext2 or ext3 filesystems to another ATA drive. By default, the copy-ata-drive program will make a bootable backup copy of the main system disk (drive hda) onto drive hde. Physical Drive Locations. and 32021003-02: In Avamar nodes with part numbers 32021003-01
hda is in drive slot 4 (the far right drive slot) hde is in drive slot 2 (the middle left drive slot) hdg is drive slot 1 (the far left drive slot)
Synopsis
copy-ata-drive [--bios=8x] [--boot=m] [--debug] [--dst=hdY] [--help] [--inplace] [--liloconfig=FILE] [--noactivate] [--nocheck] [--root=n] [--src=hdX] [--verbose]
Options
--bios=8x --boot=m --debug --dst=hdY --help --inplace --liloconfig=FILE --noactivate --nocheck --root=n --src=hdX --verbose Specifies BIOS device ID for boot. Specifies boot partition number. Default value is 1. Prints utility session information but does not actually perform the actions. Specifies destination hard drive (hdY). Default value is hde. Shows help, then exits. Sets up to boot in place (do not boot as source device). Specifies lilo configuration FILE for destination hard drive. Does not run lilo on destination drive. Does not perform source check or destination read-only bad block check. Specifies boot partition number. Default value is 5. Specifies source hard drive (hdX). Default value is hda. Provides maximum information.
218
Notes
IMPORTANT: The copy-ata-drive program must not be run on a storage node under any circumstances. The reason for this that there are typically no spare drives on a storage node. Therefore, doing so will adversely affect data stored on the server. There is currently no safeguard to prevent you from running copy-ata-drive on a storage node. Apart from --debug and --verbose, the only option normally used is --dst, which specifies an alternative destination drive. The behavior of the backup system drive is different from the original in exactly one way: all hard drives apart from hda are commented out in the backup system drive's /etc/fstab, and therefore are not mounted by default when booting on the backup system drive. Additionally, over time the contents of the backup system drive naturally might diverge from those of the original. The copy-ata-drive program performs some integrity checking on the source drive before committing the copy, primarily by doing test dumps of the data to / dev/null. An amount of time proportional to the amount of data to be copied is allotted for various read and write operations. If these operations time out, then the copy operation is deemed to have failed. The partitioning scheme of the original drive is preserved, although some partitions might grow by up to 32MB if some specialized error conditions are encountered while creating new filesystems on the destination drive. The lilo configuration of the original drive is also preserved, except for some minor modifications necessary to render the backup system drive bootable when the backup is eventually placed into the primary system slot. When using copy-ata-drive from an interactive shell, the following interactive warning prompt appears:
WARNING: This program will instantly destroy all data on destination drive hde. Proceed? y(es), n(o), q(uit/exit):
Answer y(es) in order to proceed with the copy operation. TIP: Use the following command to completely bypass the interactive warning prompt: copy-ata-drive </dev/null With the 2.4.18-3 Linux kernel and possibly with other version s, an Amax 1U node will hang during the boot process if either drive hda or hdc is not physically present. It is acceptable for either hde or hdg not to be physically present. The Linux kernel hangs while probing for drives on the motherboard's primary and secondary ATA channels. Therefore, it is best to make copies of the system disk to hde or to hdg, rather than to hdc. The --inplace option currently does not work on Amax 1U nodes because it is not known how to make the BIOS boot directly and solely from a secondary drive. Therefore, in order to make it possible for those nodes to boot on the backup system disk, you must presently accept the default behavior, which requires physically moving the backup system drive (hde by default) to the primary ATA slot (hda). AVAMAR 4.1 TECHNICAL ADDENDUM 219
Examples
This command copies hda to hde and makes hde bootable as hda: copy-ata-drive This command copies hda to hdg and makes hdg bootable as hda: copy-ata-drive --dst=hdg This command copies hda to hdc and makes it bootable in place as hdc: copy-ata-drive --dst=hdc --bios=0x81 --inplace This command copies hda to hdc and makes it bootable using a pre-defined lilo configuration file (lilo.conf.hdc): copy-ata-drive --dst=hdc --liloconfig=lilo.conf.hdc
220
copy-checkpoint
The copy-checkpoint program performs one-to-one copying of checkpoints between identically configured multi-node servers.
Synopsis
copy-checkpoint --checkpoint=cp.YYYYMMDDHHMMSS [--debug] [--degree=N] {[--destination=MODULE-1s[,MODULE-2s] | [--dprobedir=DIR]} [--dryrun] [--fakenodes] [--help] [--key=FILE] [--nodes=NODELIST] {[--source=MODULE-1s[,MODULE-2s | [--sprobedir=DIR]} [--verbose]
Options
--checkpoint= cp.YYYYMMDDHHMMSS[,...] Specifies which checkpoint to copy, where cp.YYYYMMDDHHMMSS is a valid checkpoint ID. Multiple checkpoints can be specified on the same command line; separate multiple checkpoint IDs with a comma. This option is required. --debug --degree=N --destination= MODULE-1s[,MODULE-2s] Prints utility session information but does not actually perform the actions. Specifies degree of parallelism (# of nodes at once). Specifies destination Avamar server utility nodes, where MODULE-1s and MODULE-2s are the primary and secondary utility node hostnames or IP addresses, respectively. This will be used to locate a valid probe.out file. NOTE: Most multi-node Avamar servers are deployed in a single-module configuration. Therefore, specification of a secondary utility node is usually not required. --dprobedir=DIR --dryrun --fakenodes --help --key=FILE --nodes=NODELIST Specifies directory (DIR) for existing destination Avamar server probe.out file. Runs auxiliary scripts in debug mode. Uses false module information; implies --debug. Shows help, then exits. Specifies path to destination Avamar server OpenSSH dpnid key FILE. Copies only selected source nodes. Default value is #.#.
221
copy-checkpoint COMMAND REFERENCE --source= MODULE-1s[,MODULE-2s] Specifies source Avamar server utility nodes, where MODULE-1s and MODULE-2s are the primary and secondary utility node hostnames or IP addresses, respectively. This will be used to locate a valid probe.out file. NOTE: Most multi-node Avamar servers are deployed in a single-module configuration. Therefore, specification of a secondary utility node is usually not required. --sprobedir=DIR --verbose Specifies directory (DIR) for existing source probe.out file. Provides maximum information.
Notes
IMPORTANT: You must load the dpnid OpenSSH key before running this utility. This program requires that source and destination Avamar server nodes have similar data partitioning schemes (for example, both source and destination servers have /data01 thru /data04 partitions). Prior to version 2.0.2, there was a known incompatibility between copy-checkpoint and the current probe.out file format. In order to use copy-checkpoint you had to modify the probe.out file by removing the second utility node address (the one after the comma, which was formerly the administrator node IP address). This issue has been resolved as of version 2.0.2.
Examples
Although some of these examples continue (wrap) to more than one line, all commands and options must be entered on a single command line (no line feeds or returns allowed). This example copies checkpoint cp.20030616081500 from the source Avamar server to destination Avamar server specified using the --destination option: copy-checkpoint --destination=avamar-1s,avamar-2s --checkpoint=cp.20030616081500 This multi-line example copies checkpoint cp.20030616081500 from the source Avamar server to destination Avamar server using existing stmp/probe.out and dtmp/probe.out and to define the source and destination Avamar servers, respectively: (cd stmp && export SYSPROBEDIR=. && probe dpn01 dpn02) (cd dtmp && export SYSPROBEDIR=. && probe dpn03 dpn04) copy-checkpoint --sp=stmp --dp=dtmp --checkpoint=cp.20030616081500
222
copy-checkpoint COMMAND REFERENCE This example copies checkpoint cp.20030616081500 from the source Avamar server to destination Avamar server specified using the --destination option and also specifies the path to the destination Avamar server OpenSSH dpnid key file: copy-checkpoint --destination=avamar-1s,avamar-2s --key=$HOME/.ssh/avamar-1-dpnid --checkpoint=cp.20030616081500
223
cp_cron
The cp_cron program is primarily intended to be run as a cron job in order to perform a checkpoint. In this capacity, it is typically invoked from morning_cron_run (page 323) and evening_cron_run (page 274). However, cp_cron can also be run directly to create a list of valid checkpoints in the system.
Synopsis
cp_cron [{--checkpointsfile=FILE | --nocheckpointsfile}] [--keepmin] [--nodelete] [--nolist] [--waittime=SEC]
Options
--checkpointsfile= {FILE | --nocheckpointsfile} IMPORTANT: Beginning with version 3.0.1, these options have been discontinued. The current behavior is to continuously write a list of checkpoint IDs to /usr/local/avamar/var/checkpoints.xml unless --nolist is supplied. If --checkpointsfile=FILE is supplied, a list of checkpoints is saved to this FILE. Default filename (FILE) and location is /usr/local/avamar/var/cron/checkpoints. If FILE is not supplied, output goes to STDOUT. If --nocheckpointsfile is supplied, no checkpoints list file is created. --keepmin Enables the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. This option added in version 4.0. --nodelete --nolist If supplied, checkpoints will not be deleted. If supplied, new checkpoint IDs are not added to /usr/local/avamar/var/checkpoints.xml. Specifies number of seconds (SEC) to wait for server to enter read-only mode. If the server does not enter read-only mode within this period of time, cp_cron terminates. Default value is 300. This option added in version 2.0.0.
--waittime=SEC
Notes
cp_cron must be run as user dpn.
224
cplist
The cplist program takes the XML output of avmaint lscp and displays it in a more human-readable form.
Synopsis
cplist cplist FILE [--full] [--keepmin] [--oldstyle] avmaint lscp | cplist cplist --lscp [--full] [--keepmin] [--oldstyle]
Options
--full --keepmin If supplied, all valid checkpoints are displayed. Used with --lscp option to enable the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. This option added in version 4.0. --lscp --oldstyle Lists available checkpoints. Shows all checkpoints as type hfs regardless of actual type (page 22). This option added in version 3.7.1.
Notes
The first synopsis example displays the contents of the saved file, /usr/local/avamar/var/checkpoints.xml. If this does not exist, it is created first. The second synopsis example displays the content of the specified file, assumed to be in XML format as produced by avmaint lscp. The third synopsis example illustrates its use as a filter. The fourth synopsis example creates the standard file (with all valid checkpoints if --full is specified) and then displays it.
225
Output
Output conforms to the following format: CHECKPOINT DATE CHECKED VALIDITY CHECK TYPE DELETABLE NODES REFCOUNT/NODECOUNT NUMNODES STRIPES NUMSTRIPES For example: cp.20070402185644 Mon Apr 2 11:56:44 2007 valid hfs del nodes 6/6 stripes 414 The validity attribute is either valid or invalid. The type attribute is one of the following: hfs rdc par PERCENTAGE All checks on all stripes. Reduced checks. Partial set of checks or stripes. A numerical percentage of checks complete.
Refer to Types of Checks Performed (page 22) for additional information. The REFCOUNT attribute is the number of nodes that responded to the lscp command for this checkpoint. A full checkpoint can have a lower than expected REFCOUNT if the lscp command occurred when a node was offline. The NODECOUNT attribute is the total number of online nodes that participated in the checkpoint. This can be different than the number of nodes (offline and online) on the system. A minimal checkpoint (that is, one with a node missing) will always produce a lower than expected REFCOUNT because the missing node (even if now online) will not respond for the checkpoint in question. If the checkpoint can be deleted, it is indicated with del.
226
create_newconfigs
create_newconfigs is a utility that reads the values from a config_info file (page 429) and configures the utility node in that particular module accordingly. create_newconfigs and config_info reside in the /usr/local/avamar/src directory; create_newconfigs should be run from that location.
227
cron_env_wrapper
The cron_env_wrapper program sets the environment for running Avamar cron jobs. This includes setting the PATH environment variable and starting an sshagent. IMPORTANT: This program must be run as user dpn. This program attempts to load the dpnid OpenSSH key from the dpn user .ssh directory. This program will fail if it is run as a different user.
Synopsis
cron_env_wrapper CRONJOB [--debug] [--help] [--log=FILE]
Options
CRONJOB Specifies which cron job to run. Valid Avamar cron jobs are: --debug --help --log=FILE 10minute_cron_run (page 41) 5minute_cron_run (page 40) cp_cron (page 224) evening_cron_run (page 274) gc_cron (page 278) hfscheck_cron (page 288) morning_cron_run (page 323)
Prints utility session information but does not actually perform the actions. Shows help, then exits. Logs to this FILE.
228
dbcreate_mds
IMPORTANT: Beginning with version 3.7, this program has been deprecated in favor of mds_ctl (page 316). The dbcreate_mds program creates and initializes a metadata search database on an access node. It creates the /data01/mds_data and /data01/mds_log directories, and initializes the metadata search database (mdsdb).
Synopsis
dbcreate_mds [--force] [--nocreatedb]
Options
--force The dbcreate_mds program will always warn if an existing metadata search database is detected. However, supplying -force creates and initializes a new metadata search database even if an existing metadata search database is detected. This is not the default setting. IMPORTANT: Supplying --force will completely overwrite any data in an existing metadata search database. You will not be given an opportunity to reconsider this action once the utility is invoked. This option added in version 3.6. --nocreatedb If supplied, Avamar system administrators can create the symbolic links that link metadata search to axionfs without effecting an existing metadata search database. This is not the default setting. The dbcreate_mds --nocreatedb command should be run if dbcreate_mds has already been run, and subsequently axionfs is installed on the access node. This option added in version 3.7.
Notes
This program added in version 3.5.1.
229
dbload.sh
The dbload.sh shell script loads the copy of the MCS database created by dbdump.sh.
230
dbmaint.sh
The dbmaint.sh shell script performs maintenance on various PostgreSQL databases.
Synopsis
dbmaint.sh [--db={on | off | auto}] [--dbname={emdb | mcdb | mdsdb}] [--mcs] [--script=FILE] [--server]
Options
--db= {on | off | auto} Controls database state after maintenance is performed. One of the following: auto Leaves database in same state as when maintenance started. This is the default setting. Shuts down database on exit. Leaves database running on exit.
Specifies which database to process. One of the following: emdb mcdb mdsdb Process Avamar Enterprise Manager server database. Process MCS database. Process metadata search database.
--mcs
Allows dbmaint.sh to run while MCS is running, otherwise dbmaint.sh can only be run while the MCS is stopped. If FILE is a .sql file, then it is processed as a database script. Otherwise, it is processed as a shell script. No script is executed if this option is not specified. .sql files are located in lib/sql; shell scripts are located in /usr/local/avamar/bin.
--script=FILE
--server
231
Scripts
The dbmaint.sh shell script accepts one (and only one) of the following script files with the --script=FILE option: db_count.sql dbcreate.sql dbcreateuser.sql dbdropviews.sql dbdump.sh dbinitdatasets.sql dbinitevents.sql dbinitgroups.sql dbinitretpolicies.sql dbinitschedules.sql dbmaint.sql dbupgrade.sql Returns total number of records in all MCS database tables. Creates the MCS database tables. Creates MCS database users. Removes MCS database views. Creates a portable copy of the MCS database. Loads default dataset data. Loads default event definitions. Loads default group definitions. Loads default retention policy definitions. Loads default schedule definitions. Frees unused hard drive space by running an SQL vacuum command against the mcdb. IMPORTANT: Beginning with version 1.2.1, support for this SQL file has been discontinued. Updates database tables to current MCS database schema. dbviews.sql ec_update.sql emupgrade.cidlength.sql Creates MCS database views. Forces an upgrade of the event catalog by resetting the version numbers. Performs field upgrade of version 3.5.0 thru 3.7.0 EMS databases in order to resolve a client ID length issue. This script added in version 3.7.1. show_versions.sql Returns schema version.
Notes
This shell script added in version 1.2. It supersedes and obsoletes mcdbmaint.sh (page 302) and mcdbsql.sh (page 303).
232
dbpurge.sh
The dbpurge.sh shell script manually purges unused data from and reclaims tablespace in the MCS PostgreSQL database (mcdb). IMPORTANT: You should only purge data if it is no longer needed for future report generation. If the data is needed, extract it to another storage device (database, archive) then purge it from the MCS database. Operational data tables typically do not grow very large in size and only grow in direct proportion to the management of clients, groups, schedules, retention policies and event profiles. However, other report data tables can grow very large over time and should be purged to avoid accumulation of old data: The activities table (mcdb.activities) increases in size after each backup and restore request is processed The events table (mcdb.events) can increase in size every second, depending on system loading and system events The node utilization (mcdb.sv_node_util) and node capacity (mcdb.sv_node_space) tables increase in size every 10 minutes
Synopsis
dbpurge.sh --from=YYYY-MM-DD --to=YYYY-MM-DD {--all | --v_activities | --v_events | --v_node_space | --v_node_util}
Options
--from=YYYY-MM-DD --to=YYYY-MM-DD --all Used with --to=YYYY-MM-DD to define a range of dates for the purge operation. This option is required. Used with --from=YYYY-MM-DD to define a range of dates for the purge operation. This option is required. Purges the activities (mcdb.activities), events (mcdb.events), node capacity (mcdb.sv_node_space) and node utilization (mcdb.sv_node_util) tables. Purges the activities table (mcdb.activities). Purges the events table (mcdb.events). Purges the node capacity table (mcdb.sv_node_space). Purges the node utilization table (mcdb.sv_node_util).
233
Examples
This command returns the size (count) of the tables distributed by dates, which can be used to determine which tables and date ranges to purge: /usr/local/avamar/bin/dbmaint.sh --mcs --script=db_v_count.sql This command purges the activities (mcdb.activities), events (mcdb.events), node capacity (mcdb.sv_node_space) and node utilization (mcdb.sv_node_util) tables of all data from September 16, 2003 (2003-09-16) to September 20, 2003 (200309-20): dbpurge.sh --from=2003-09-16 --to=2003-09-20 --all
234
dbUpdateActivitiesExt.pl
The dbUpdateActivitiesExt.pl Perl script updates the mcdb activities_ext table. An Avamar server will require updating by running this script if it was upgraded to version 3.5.0 or if after upgrading to version 3.5.0, the Avamar server is rolled back to a point prior to the 3.5.0 upgrade. The dbUpdateActivitiesExt.pl program is located in /usr/local/avamar/lib/sql. This Perl script added in version 3.5.
235
delete-snapups
The delete-snapups is a program that writes a script to stdout, which when run, deletes any backups stored on the Avamar server with a creation date prior to the specified date.
Synopsis
delete-snapups [--after=DATE] [--before=DATE] [--domain=CLIENT-DOMAIN] [--help] [--include_mc_backups] [CLIENT-PATH ...]
Options
--after=DATE If supplied, all backups created after this date are eligible to be deleted. Default DATE setting is June 1, 1999 00:00:00. If supplied, all backups created before this date are eligible to be deleted. Default DATE setting is two weeks ago. If supplied, all clients within the specified CLIENTDOMAIN are processed. Shows help, then exits. If supplied, all clients within the special MC_BACKUPS domain are also processed. Because MC_BACKUPS is a special system domain, the default behavior is to not process these clients unless this option is explicitly supplied.
--before=DATE
Client Path
CLIENT-PATH Specifies a valid Avamar client path (for example, /clients/MyClient). Additional client paths can be delimited with white space. IMPORTANT: Client paths must always terminate with a valid Avamar client name. You cannot specify a client domain in this manner. Use --domain when you want to process all clients within a particular domain. If not supplied, all clients are processed. This is the default setting.
236
Notes
IMPORTANT: This utility deletes backups but does not update the MCS database. This will cause the total bytes used value, which is used by the server licensing mechanism, to be incorrect. Therefore, EMC strongly suggests that you only use this utility if instructed to do so by EMC Technical Support. Contact EMC Technical Support for additional information. After deleting backups with this utility, you should immediately run dbUpdateactivitiesExt.pl (page 235) to update the MCS database. If you do not do this, your servers total bytes used value will be incorrect, which will adversely affect your server licensing (you will be deleting backups, which would normally free up additional storage capacity, but the licensing mechanism will be unaware that you have done so). The DATE specifier can be any date string acceptable to date(1). For GNU date(1), which on Linux is /bin/date, DATE can be just about any common date string, including such phrases as 2 weeks ago. This program added in version 3.0.
Examples
To create a default output script (one that will delete any backup stored /clients with a creation date older than two weeks from todays date) with the user-defined name del-old-backups, enter: delete-snapups /clients > del-old-backups To restrict program execution to specific client area (/clients/engineering in this example) and change the time period to three weeks ago, enter the following on a single command line: delete-snapups --before='3 weeks ago' /clients/engineering > del-old-backups After creating a script with the desired backup deletion parameters, enter the following to actually delete the backups from the Avamar server: /bin/sh -x ./del-old-backups
237
dpn.pm
The dpn.pm file is a Perl module used by many of the Perl scripts in the dpnutils package.
dpncron.pm
The dpncron.pm file is a Perl module used by cp_cron (page 224), gc_cron (page 278) and hfscheck_cron (page 288).
238
dpnctl
The dpnctl program performs two functions: Implements unattended automated shutdowns and restarts of single-node servers Simplifies shutdowns and restarts on all Avamar servers
Synopsis
dpnctl {{disable | enable | help | start | status | stop} [SUBSYSTEM]} | uninstall} [--config=FILE] [--debug] [--ems_password=PASSWORD] [--ems_shutdown] [--ems_user=ID] [--event_notification_timeout=SEC] [--force_ems_restore] [--force_mcs_restore] [--force_rollback] [--force_version_record_update] [--help] [--hfsport=PORT] [--ignore_lock] [--ignore_node_type] [--mcs_password=PASSWORD] [--mcs_user=USER-ID] [--nointeractive_questions] [--now] [--override_interactive_update_requirement] [--probedir=PATH] [--require_clean_gsan] [--update] [--use_status_dpn] [--verbose] [--xml]
Commands
One (and only one) of the following commands must be supplied on each command line: disable enable help Disables automated server shutdowns and restarts. Enables automated server shutdowns and restarts. Shows help, then exits. Same as supplying the --help option.
239
dpnctl COMMAND REFERENCE start [SUBSYSTEM] Starts server subsystems. Valid SUBSYSTEM codes are: all Start all server subsystems (axionfs, ems, gsan, mcs and scheduler (sched)). This is the default setting. Only start Avamar virtual filesystem (axionfs) subsystem. This subsystem added in version 3.5. ems Only start EMS subsystem. This subsystem added in version 3.5. gsan maint Only start storage (gsan) subsystem. Only start maintenance subsystem. This subsystem added in version 3.5. mcs sched status [SUBSYSTEM] Only start MCS and scheduler (sched) subsystems. Only start scheduler (sched) subsystem.
axionfs
Shows status of server subsystems. Valid SUBSYSTEM codes are: all Show status of all server subsystems (axionfs, ems, gsan, mcs and scheduler (sched)). This is the default setting. Only show status of Avamar virtual filesystem (axionfs) subsystem. This subsystem added in version 3.5. ems Only show status of EMS subsystem. This subsystem added in version 3.5. enable gsan maint Show whether or not automated server shutdowns and restarts is enabled. Only show status of storage (gsan) subsystem. Only show status of maintenance subsystem. This subsystem added in version 3.5. mcs sched Only show status of MCS and scheduler (sched) subsystems. Only show status of scheduler (sched) subsystem.
axionfs
240
dpnctl COMMAND REFERENCE stop [SUBSYSTEM] Stops server subsystems. Valid SUBSYSTEM codes are: all Stop all server subsystems (axionfs, ems, gsan, mcs and scheduler (sched)). This is the default setting. Only stop Avamar virtual filesystem (axionfs) subsystem. This subsystem added in version 3.5. ems Only stop EMS subsystem. This subsystem added in version 3.5. gsan maint Only stop storage (gsan) subsystem. Only stop maintenance subsystem. This subsystem added in version 3.5. mcs sched uninstall Only stop MCS and scheduler (sched) subsystems. Only stop scheduler (sched) subsystem.
axionfs
Options
--config=FILE Specifies configuration FILE location. Supplying --config=- (equal sign and dash) reads XML configuration data from stdin. --debug --event_notification_timeout=SEC Prints program session information but does not actually perform the actions. Specifies timeout in seconds (SEC) for event notification. Specify zero (0) for no timeout. Default value is 1800 seconds. Specifies alternative EMS password. This option added in version 3.7.
--ems_password=PASSWORD
241
dpnctl COMMAND REFERENCE --ems_shutdown Used with stop to force a shutdown of the EMS. This is the default condition. If neither --ems_shutdown nor --noems_shutdown is supplied, and dpnctl is operating in an interactive context (unattended automated server shutdowns are not enabled), an interactive prompt will appear, asking the user if they want to shut down this instance of the EMS. This option added in version 3.5. --ems_user=ID Specifies alternative EMS user ID. This option added in version 3.7. --force_ems_restore Used with start to force an EMS restore before restarting. This option added in version 3.7. --force_mcs_restore --force_rollback --force_version_record_update Used with start to force an MCS restore before restarting. Used with start to force a rollback before restarting. Used with start to force an update of the version record. This option added in version 4.0. --help --hfsport=PORT Shows help, then exits. Same as supplying the help command. Specifies data PORT to use for communications between the Hash Filesystem (HFS) and the MCS when MCS restores are performed. Default data port is 27000. Ignores lock file on startup. Bypasses node type checking. Default behavior is to exit without comment when dpnctl is run on a multi-node server in a non-interactive context. Specifies MCS password. Refer to Passwords (page 245) for additional information. Specifies MCS user ID. Default value is root.
--ignore_lock --ignore_node_type
--mcs_password=PASSWORD
--mcs_user=ID
242
dpnctl COMMAND REFERENCE --nointeractive_questions If supplied, dpnctl behaves as though it were executing in a non-interactive context. This is not the default setting. This option added in version 3.7. --now Used with stop command to perform an immediate system shutdown without waiting for completion of active backup sessions. If supplied, allows dpnctl start to continue when a server software update is indicated in a noninteractive context. This option added in version 4.1. --probedir=PATH --require_clean_gsan Specifies parent directory for probe.out. If suppled, causes dpnctl start to exit with an error message if it encounters data on the server (that is, if server is not clean). This option added in version 4.1.
--override_interactive_update_requirement
243
dpnctl COMMAND REFERENCE --update Supplying --update forces dpnctl to operate in an update context; supplying --noupdate inhibits dpnctl from operating in an update context. Beginning with version 3.5, if a newer version dpnmcs software package is detected during a 3.5 or later server software upgrade, then dpnctl will determine that this is an update context and propagate the correct version gsan binary to all storage nodes. In an update context, dpnctl might invoke one or more of the following commands: mcserver.sh --update (page 309) (unless an MCS data restore has occurred) avsetup_mccli (page 163) avsetup_webstart (page 169) avsetup_ems (page 162) emserver.sh --init (page 268) (once only) website create-cfg (page 420) website init (page 420) website restart (page 420) service lm restart (page 296) avmaint config lmaddr=UTILITY-NODE-IP-ADDR (page 76) --use_status_dpn Used with status command to force use of older status.dpn (page 396) utility instead of opstatus.dpn (page 326). This option added in version 3.7. --verbose --xml Provides maximum information. Used with status command to format output in XML.
Notes
This program added in version 3.0; axionfs, ems and maint subsystems added in version 3.5. This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option. The dpnctl program does not presently handle the first-time-ever startup case. When starting up a system for the first time, do not use dpnctl but instead, use the documented procedures for first-time startup. If the dpn user's crontab is not installed, dpnctl status might still report that maintenance cron jobs are enabled (that they are not suspended). If maintenance cron jobs are not running, check that the dpn user's crontab is installed. If garbage collection is active while applying dpnctl stop, then the read-only status of gsan (0000+0000+0000) might cause complaints during administrator server shutdown.
244
Files Used
/usr/local/avamar/etc/usersettings.cfg /usr/local/avamar/var/dpn-auto-start.DISABLE Client settings file. Inhibits automatic startup when present (used by enable and disable commands). Lock file. Log file (when --log is used). This log file is automatically rotated when it reaches 1MB in size; up to eight rotated dpnctl log files are retained in /usr/local/avamar/var/log/. probe.out Produced by running probe mod1 [mod2 ...].
/usr/local/avamar/var/dpnctl.lck /usr/local/avamar/var/log/dpnctl.log
Passwords
For default MCS user root, the client settings file is consulted if no password is supplied by way of command line or configuration file. MCS user and password information must be available in order to perform start, stop, or status operations on schedules; or to perform rollback and restart operations.
245
dpnctl COMMAND REFERENCE password in the client settings file is invalid for a particular checkpoint. If necessary, intervene manually to correct the client settings file before restarting. For an MCS restore operation, the client settings file is examined if no --hfsport=PORT option is supplied by way of command line or configuration document. Default data port for communications between the Hash Filesystem (HFS) and the MCS is 27000.
If the storage server (formerly known as GSAN) is running, then a <gsan name="SYSTEM-NAME"/> element is present as a child element under <identification>. <aggregate status= up> is present if both the storage server (formerly known as GSAN) and MCS are running; <aggregate status= down> is present otherwise. <upgradeable status= yes> is present if both the storage server (formerly known as GSAN) and MCS return statuses of ready and up, respectively; <upgradeable status= no reason=EXPLANATORY-TEXT> is present otherwise. The determination as to whether the storage server or MCS is up does not take into account stripe status or a read-only system condition. The storage server is defined to be up if opstatus.dpn (page 326) exits with status 0 and without printing an ERROR or FATAL message. The MCS is defined to be up if mcserver.sh --ping (page 309) exits with status 0 and without printing an ERROR or FATAL message.
246
#REQUIRED> #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED #IMPLIED> #REQUIRED #REQUIRED #REQUIRED> #REQUIRED #REQUIRED #REQUIRED>
<!ELEMENT <!ATTLIST
<!ELEMENT <!ATTLIST
Pass-Through Feature
The dpnctl pass-through feature allows you to persistently customize which options are passed through to the various other programs dpnctl invokes. In order to use the pass-through feature, you must: 1. Create a configuration file, containing your custom pass-through settings. 2. Invoke dpnctl with the --config=FILE option, which will force dpnctl to read that configuration FILE and use those pass-through settings. Configuration File. In order to create a valid configuration file, use a Unix text editor such as vi or Emacs to create a plain text file with a meaningful name in a convenient directory. /usr/local/avamar/dpnctl.cnf will be used as an example configuration filename and path for the remainder of this topic. For each pass-though command, you must define an operational context (start, stop, restore or update) and a valid program that dpnctl is known to invoke in that operational context.
247
dpnctl COMMAND REFERENCE The following table lists valid programs that can have pass-through options defined for them in each operational context:
CONTEXT PROGRAM ACTUAL COMMAND
start
emserver.sh --start [PASS-THROUGH-OPTIONS] mcserver.sh --start [PASS-THROUGH-OPTIONS] restart.dpn [PASS-THROUGH-OPTIONS] rollback.dpn --cptag=CP-ID [PASS-THROUGH-OPTIONS] emserver.sh --stop [PASS-THROUGH-OPTIONS] mcserver.sh --stop [PASS-THROUGH-OPTIONS] shutdown.dpn [PASS-THROUGH-OPTIONS] mcserver.sh --restore --id=USER-ID --hfsport=PORT --hfsaddr=IP-ADDR --password=PASSWORD [PASS-THROUGH-OPTIONS] emserver.sh --restore --id=USER-ID --hfsport=PORT --hfsaddr=IP-ADDR --password=PASSWORD [PASS-THROUGH-OPTIONS] mcserver.sh --update [PASS-THROUGH-OPTIONS]
stop
restore
mcserver.sh
emserver.sh
update
mcserver.sh
Refer to emserver.sh (page 268), mcserver.sh (page 309), restart.dpn (page 364), rollback.dpn (page 370) and shutdown.dpn (page 378) for comprehensive information about each programs available command line options. The structure of the configuration file is XML. <passthrough> elements are used to define operational context, program and one or more command line options that will be passed to that program in that operational context. Practical Example. Consider the following example entry in /usr/local/avamar/dpnctl.cnf:
<dpnctl> <passthrough program="restart.dpn" context="start" options="--delay"/>
</dpnctl>
This configuration file entry passes through the --delay command line option to the restart.dpn program each time dpnctl start is invoked with the --config=/usr/local/avamar/dpnctl.cnf option. Invoking dpnctl start without the --config=/usr/local/avamar/dpnctl.cnf option bypasses any pass-through settings stored in that configuration file.
248
249
dpnfsctl
The dpnfsctl program directly controls the Avamar File System (AvFS). IMPORTANT: Documentation for this program is provided for reference purposes only. The dpnfsctl program is typically invoked programmatically by dpnctl (page 239) and should not be run directly from the command line.
Synopsis
dpnfsctl {help | start | status | stop} [--debug] [--help] [--verbose]
Commands
One (and only one) of the following commands must be supplied on each command line: help start status stop Shows help, then exits. Same as supplying the --help option. Starts Avamar File System (AvFS). Shows status of Avamar File System (AvFS). Stops Avamar File System (AvFS).
Options
--debug --help --verbose Prints program session information but does not actually perform the actions. Shows help, then exits. Same as supplying the help command. Provides maximum information.
Notes
This program added in version 3.5.
250
dpnnetutil
The dpnnetutil program is an interactive utility that assists in reconfiguring Avamar server network configurations. Both multi-node and single node servers are supported.
Synopsis
dpnnetutil [--broadcast] [--help]
Options
--broadcast If supplied, dpnnetutil performs a broadcast ping on eth0, the results of which provide a pool of initial candidate node IP addresses. The first 38 nodes to respond within a short window of time are added to the pool, excluding the first and last addresses of the subnet. Explicitly supplying --broadcast option implies that the person running this utility knows that this is a multi-node server configuration. Therefore, interactive prompting for server type is bypassed. --help Shows help, then exits.
Notes
This program added in version 3.7.1. IMPORTANT: dpnnetutil must be run as root. It also requires the dpnid OpenSSH key in order to perform operations as root on a multi-node system.
Log Files
All log files are located in /usr/local/avamar/var/log/.
LOCATION FILENAME DESCRIPTION
utility node
dpnnetutil.log dpnnetutilbg.out
This is a detailed master log file for dpnnetutil. This file contains a progress summary for the non-interactive background job, dpnnetutilbg.
251
all nodes
dpnnetutilbgauxstdout-stderr.log
This file contains a progress summary for the non-interactive per-node background helper job, dpnnetutilbgaux, which is responsible for making configuration changes on each node. This file contains detailed log information from the non-interactive per-node background helper job, dpnnetutilbgaux.
dpnnetutilbgaux.log
Files Modified
FILENAME UPDATE TYPE ATTRIBUTE(S) UPDATED DESCRIPTION
/etc/hosts
Full
Avamar server node address/hostname mappings. Hostname. Default network gateway IP address. Enable networking. Network interface name. Boot protocol. IP address for interface eth0. Network mask for interface eth0. Enable network interface at boot time. Primary and secondary DNS resolving name server addresses. Search domain for unqualified hostnames. Name of parent domain.
/etc/sysconfig/ network
Partial
/etc/sysconfig/ networkscripts/ifcfgeth0
Partial
/etc/ resolv.conf
Full
nameserver IP-ADDRESS
search DOMAIN-NAME
domain IP-ADDRESS
252
FILENAME
DESCRIPTION
Partial
node type=NODE-TYPE
Node type (if one of the supported node types). Addresses for node types supported by the current probe.out file format.
Full
253
(a) Can dpnnetutil programmatically find any available IP addresses, which might be assignable to storage nodes? No: Go to step (c). Yes: Go to step (b) (b) Prompt user to eliminate IP addresses that should not be assigned to storage nodes: Please review storage node IP addresses below and mark (set) any entries to be removed from the list. Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address marked for removal is removed from the list of available IP addresses that can be assigned to storage nodes. (c) Prompt user to manually enter any known storage node IP addresses that might not have been automatically detected: Please enter the old (current) IP address of a storage node. Leave the field empty when no more storage node addresses remain to enter. This step loops until all known storage node IP addressees have been entered, then repeat step b one last time, leaving the input field empty. (d) Prompt for user confirmation: Please review storage node IP addresses below and mark (set) any entries to be removed from the list. (e) Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address not marked for removal is removed from the list of available IP addresses that can be assigned to storage nodes. (f) Prompt for user confirmation: Done entering storage node IP addresses? No: Return to step 2.(b). Yes: Go to step 3. 3. Enumerate available accelerator node IP addresses. NOTE: Accelerator nodes are completely optional.
254
dpnnetutil COMMAND REFERENCE (a) Prompt user to eliminate IP addresses that should not be assigned to accelerator nodes: Please review accelerator node IP addresses below and mark (set) any entries to be removed from the list. Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address marked for removal is removed from the list of available IP addresses that can be assigned to accelerator nodes. (b) Prompt user to manually enter any known accelerator node IP addresses that might not have been automatically detected: Please enter the old (current) IP address of a accelerator node. Leave the field empty when no more accelerator node addresses remain to enter. This step loops until all known accelerator node IP addressees have been entered, then repeat step b one last time, leaving the input field empty. (c) Prompt for user confirmation: Please review accelerator node IP addresses below and mark (set) any entries to be removed from the list. (d) Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address not marked for removal is removed from the list of available IP addresses that can be assigned to accelerator nodes. (e) Prompt for user confirmation: Done entering accelerator node IP addresses? No: Return to step 3.(b). Yes: Go to step 4. 4. Enumerate available access node IP addresses. NOTE: Access nodes are completely optional. (a) Prompt user to eliminate IP addresses that should not be assigned to access nodes: Please review access node IP addresses below and mark (set) any entries to be removed from the list. Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address marked for removal is removed from the list of available IP addresses that can be assigned to access nodes. (b) Prompt user to manually enter any known access node IP addresses that might not have been automatically detected: Please enter the old (current) IP address of a access node. Leave the field empty when no more access node addresses remain to enter. This step loops until all known access node IP addressees have been entered, then repeat step b one last time, leaving the input field empty. (c) Prompt for user confirmation: Please review access node IP addresses below and mark (set) any entries to be removed from the list. AVAMAR 4.1 TECHNICAL ADDENDUM 255
dpnnetutil COMMAND REFERENCE (d) Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address not marked for removal is removed from the list of available IP addresses that can be assigned to access nodes. (e) Prompt for user confirmation: Done entering access node IP addresses? No: Return to step 4.(b). Yes: Go to step 5. 5. Enumerate available spare node IP addresses. NOTE: Spare nodes are completely optional. (a) Prompt user to eliminate IP addresses that should not be assigned to spare nodes: Please review spare node IP addresses below and mark (set) any entries to be removed from the list. Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address marked for removal is removed from the list of available IP addresses that can be assigned to spare nodes. (b) Prompt user to manually enter any known spare node IP addresses that might not have been automatically detected: Please enter the old (current) IP address of a spare node. Leave the field empty when no more spare node addresses remain to enter. This step loops until all known spare node IP addressees have been entered, then repeat step b one last time, leaving the input field empty. (c) Prompt for user confirmation: Please review spare node IP addresses below and mark (set) any entries to be removed from the list. (d) Review the list of IP address, set (mark) any entries that should be removed from this list and choose OK. Each address not marked for removal is removed from the list of available IP addresses that can be assigned to spare nodes. (e) Prompt for user confirmation: Done entering spare node IP addresses? No: Return to step 5.(b). Yes: Go to step 6. 6. Assign utility node IP address and hostname. (a) Prompt for new utility node IP address: Please enter the new IP address for utility node OLD-IP-ADDRESS. (b) Prompt for new utility node hostname: Please enter the new host name (without the dotted domain) for the utility node whose old IP address is OLD-IP-ADDRESS.
256
dpnnetutil COMMAND REFERENCE 7. Assign storage node IP addresses and hostnames. (a) Prompt for new storage node IP address: Please enter the new IP address for storage node OLD-IP-ADDRESS. (b) Prompt for new storage node hostname: Please enter the new host name (without the dotted domain) for the storage node whose old IP address is OLD-IP-ADDRESS. Steps (a) thru (b) loop until all storage node IP addresses and hostnames have been assigned. 8. Assign accelerator node IP addresses and hostnames. NOTE: This processing only occurs if you enumerated one or more accelerator node IP addresses in step 3. (a) Prompt for new accelerator node IP address: Please enter the new IP address for accelerator node OLD-IP-ADDRESS. (b) Prompt for new accelerator node hostname: Please enter the new host name (without the dotted domain) for the accelerator node whose old IP address is OLD-IP-ADDRESS. Steps (a) thru (b) loop until all accelerator node IP addresses and hostnames have been assigned. 9. Assign access node IP addresses and hostnames. NOTE: This processing only occurs if you enumerated one or more access node IP addresses in step 4. (a) Prompt for new access node IP address: Please enter the new IP address for access node OLD-IP-ADDRESS. (b) Prompt for new access node hostname: Please enter the new host name (without the dotted domain) for the access node whose old IP address is OLD-IP-ADDRESS. Steps (a) thru (b) loop until all access node IP addresses and hostnames have been assigned. 10. Assign spare node IP addresses and hostnames. NOTE: This processing only occurs if you enumerated one or more spare node IP addresses in step 5. (a) Prompt for new spare node IP address: Please enter the new IP address for spare node OLD-IP-ADDRESS. (b) Prompt for new spare node hostname: Please enter the new host name (without the dotted domain) for the spare node whose old IP address is OLD-IP-ADDRESS. Steps (a) thru (b) loop until all spare node IP addresses and hostnames have been assigned.
257
dpnnetutil COMMAND REFERENCE 11. Prompt for default parent domain: Please enter the name of the parent domain to be used for all nodes. You may modify this later for individual nodes. 12. Prompt for primary DNS resolver IP address: Please enter the IP address of the primary DNS resolver. 13. Prompt for secondary DNS resolver IP address: Please enter the IP address of the secondary DNS resolver. NOTE: This processing only occurs if you entered a primary DNS resolver IP address in step 12. 14. Prompt for network mask: Please enter the network mask to be used for network interface eth0 on all nodes. You may modify this later for individual nodes. 15. Prompt for default network gateway IP address: Please enter the default network gateway address to be used for all nodes. You may modify this later for individual nodes. 16. Enter the correct IP address for the default network gateway all nodes on this server will be using and choose OK. 17. Ask if custom network masks are required: Do any of the nodes require a network mask other than GLOBAL-DEFAULT-NET-MASK? YES: Go to step 17.(a). NO: Go to step 18. (a) Prompt for custom network mask: Please enter the network mask to be used for network interface eth0 on HOSTNAME [NEW-IPADDRESS]. Leave this blank if you are done entering custom mask values. (b) Enter the correct network mask for the eth0 NIC on this node and choose OK. Steps (a) thru (b) loop until all custom network masks have been assigned. 18. Ask if custom default gateway settings are required: Do any of the nodes require a default gateway other than GLOBAL-DEFAULT-GATEWAY? YES: Go to step 18.(a). NO: Go to step 19. (a) Prompt for custom default gateway settings: Please enter the default gateway to be used on HOSTNAME [NEW-IP-ADDRESS]. Leave this blank if you are done entering custom default gateways. (b) Enter the correct IP address for the network gateway this node will be using and choose OK. Steps (a) thru (b) loop until all custom default gateway settings have been assigned.
258
dpnnetutil COMMAND REFERENCE 19. Ask if custom parent domain settings are required: Do any of the nodes require a parent domain other than GLOBAL-DEFAULT-PARENTDOMAIN? YES: Go to step 19.(a). NO: Go to step 20. (a) Prompt for custom parent domain settings: Please enter the parent domain to be used on HOSTNAME [NEW-IP-ADDRESS]. Leave this blank if you are done entering custom parent domains. (b) Enter the correct DNS domain name for this node and choose OK. Steps (a) thru (b) loop until all custom parent domain settings have been assigned. 20. Ask if custom primary domain name server settings are required: Do any of the nodes require a primary domain name server other than GLOBAL-DEFAULT-PRIMARY-NAME-SERVER? YES: Go to step 20.(a). NO: Go to step 21. (a) Prompt for custom primary domain name server settings: Please enter the IP address of the primary name server to be used for HOSTNAME [NEW-IP-ADDRESS]. Leave this blank if you are done entering custom primary name servers. (b) Enter the correct IP address for the primary DNS resolver used by this node and choose OK. Steps (a) thru (b) loop until all custom primary domain name server settings have been assigned. 21. Ask if custom secondary domain name server settings are required: Do any of the nodes require a secondary domain name server other than GLOBAL-DEFAULT-SECONDARY-NAME-SERVER? YES: Go to step 21.(a). NO: Go to step 22. (a) Prompt for custom secondary domain name server settings: Please enter the IP address of the primary name server to be used for HOSTNAME [NEW-IP-ADDRESS]. Leave this blank if you are done entering custom secondary name servers. (b) Enter the correct IP address for the secondary DNS resolver used by this node and choose OK. Steps (a) thru (b) loop until all custom secondary domain name server settings have been assigned. 22. Prompt for user confirmation: Accept settings and proceed to fix up the network configurations? Yes: Complete. No: Return to step 2.(a).
259
dpnsummary
The dpnsummary program returns summary information about data stored in the Avamar server and statistical data for each client backup.
Synopsis
dpnsummary [--days=NUM] [--help] [--hfsaddr=AVAMARSERVER] [--id=USER@AUTH] [--password=PASSWORD]
Options
--days=NUM Shows information for this past number (NUM) of days. If not supplied, dpnsummary shows information since the server was last started. --help --hfsaddr=AVAMARSERVER Shows full online help (commands, options and full descriptions) for this utility, then exits. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). This value is typically recorded in a configuration file. It can also be set by way of an environment variable or on the command line with each transaction. Default address is localhost (127.0.0.1). --id=USER@AUTH Authenticate as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --password=PASSWORD PASSWORD for the --id=USER@AUTH account.
Files
Files created by dpnsummary are: AVAMARSERVER-YYYY-MMM-DD-HHMM-1.log AVAMARSERVER-YYYY-MMM-DD-HHMM-1.raw AVAMARSERVER-YYYY-MMM-DD-HHMM-1.trim AVAMARSERVER-YYYY-MMM-DD-HHMM-1.csv Where AVAMARSERVER is the Avamar server hostname as defined in DNS and YYYY-MMM-DD-HHMM is a time stamp comprised of a four-digit year (YYYY), three-character month (MMM), two-digit day of the month (DD) and a four-digit time of day (HHMM). .log File. The .log file contains dpnsummary utility session (application log) information.
260
dpnsummary COMMAND REFERENCE .raw and .trim Files. The remaining files contain the actual data collected by dpnsummary. The .raw file is the largest and most verbose. The .trim file contains a subset of that information which is derived by removing periodic backup status messages, which typically occur every 30 seconds. .csv File. The .csv file presents the data in a format that can be readily imported into spreadsheet applications for analysis and presentation. Each .csv file record (line) is a specific client backup. The following table describes the various .csv file attributes (columns) in the order they appear from left to right:
ATTRIBUTE DESCRIPTION
Client hostname as defined in DNS. This is the client that is backing up data to the Avamar server. Unix time stamp for this operation. Client operating system. Date and time this backup was initiated. Starting location (top level directory) in the client filesystem for this backup. Duration (in seconds) of that client backup. Total number of files examined during this backup. Total number of modified files stored during this backup. Total number of modified file bytes not sent to the server due to data de-duplication (this unique sub-file chunk already exists on the server). For example, if during a backup, there was only one changed file that was 100 MB in size, this would be indicated with NumModFiles=1. Furthermore, if because of sub-file level data de-duplication, only 20 MB of the 100 MB was actually sent to the server for storage, then ModNotSent = 80 MB x 1024 KB/ MB x 1024 B/KB, and ModSent = 20 MB x 1024 KB/MB x 1024 B/KB. This value is of interest in evaluating data de-duplication efficiency because this provides a direct comparison between how much data is sent across the network during an Avamar backup versus how much data is backed up using standard tape incremental backup. More generally, it provides a direct indication of how efficient Avamars sub-file data de-duplication is versus backing up entire changed files.
Total number of modified file bytes sent to the server during this backup. Total number of file bytes examined during this backup. Data de-duplication percentage during this backup. This is derived by way of the following formula: 100 - (TotalBytes/ModSent)
261
Overhead
Number of bytes COMPOSITE and DIRELEMs used to store data. This is the amount of non-file data sent across from the client to the server, and includes such things as indexing information, requests from the client to the server regarding presence of specific data chunks, ACLs, directory information and message headers. On any relatively active filesystem, this is generally a very small percentage of the file data that is sent to the server.
WorkOrderID
Unique identifier for that backup. For scheduled backups, the backup name is created as follows: SCHEDULE_NAME-GROUP_NAME-UNIX_TIME_IN_MS For on-demand backups initiated by way of the Policy window Back Up Group Now command, the backup name is created as follows: GROUP_NAME-UNIX_TIME_IN_MS
ClientVer Operation
Avamar client software version. One of the following: avtarbackup avtarrestore replsrc repldest
Status
Final status of that client backup. One of the following: DONE - backup completed, even though it might have had errors, such as errors associated with open files, and so forth RUNNING - backup is in progress TERMINATED - backup did not complete either because it failed or because the backup was cancelled A backup with a status of TERMINATED is immediately eligible to have the data cleaned up on the server by way of garbage collection and there will be no record of the backup in either the Backup Management or Backup and Restore windows.
SessionID
262
dpn-time-config
The dpn-time-config program generates ntp.conf or step-tickers files for Avamar nodes. IMPORTANT: Documentation for this script is provided for reference purposes only. The dpn-time-config program is typically invoked programmatically by mktime (page 318) and should not be run directly from the command line.
Synopsis
dpn-time-config here=SUBNET peer=SUBNET [--help] [--ips] [server=IP-ADDR] [type={data | mcs | services}] [--verbose]
Options
here=SUBNET peer=SUBNET --help --ips server=IP-ADDR Specify this module's subnet (for example, 10.0.42.0/ 24). Specify peer module's subnet (for example, 10.0.43.0/ 24). Shows help, then exits. Print only the time server IP addresses (for step-tickers). IPv4 address or time zone (selects public servers). Valid time zones are: type=TYPE alaska:U.S. Alaska Time (UTC -9, -8 DST) pacific:U.S. Pacific Time (UTC -8, -7 DST) mountain:U.S. Mountain Time (UTC -7, -6 DST) central:U.S. Central Time (UTC -6, -5 DST) eastern:U.S. Eastern Time (UTC -5, -4 DST)
Node TYPE. Possible values are: data (configure for storage nodes) mcs (configure for utility node) services (configure for utility node)
--verbose
Notes
In subnet specifications, both net widths and net masks are accepted (for example, 10.0.42.0/24 or 10.0.42.0/255.255.255.0) Configuration philosophies: Do specify customer premises time servers Only specify public time servers if fewer than two customer premises time servers are available Never add public time servers to storage nodes AVAMAR 4.1 TECHNICAL ADDENDUM 263
Examples
The following examples use this bash shell script to generate a set of ntp.conf files and step-tickers files suitable for each different type of node in each module:
#!/bin/bash prog=dpn-time-config mkfiles() { for type in services mcs data ; do ${prog} type=$type $2 >ntp.conf.$type-node-$1 ${prog} -ips type=$type $2 >step-tickers.$type-node-$1 done }
Example 2 No customer premises time servers; use public time servers in the U.S. Pacific Time Zone:
SERVERS="server=pacific" mkfiles "dpn00" "$SERVERS here=10.0.42.0/24 peer=10.0.43.0/24" mkfiles "dpn01" "$SERVERS here=10.0.43.0/24 peer=10.0.42.0/24"
Example 3 No customer premises time servers and no public time servers (use internal servers only):
mkfiles "dpn00" "here=10.0.42.0/24 peer=10.0.43.0/24" mkfiles "dpn01" "here=10.0.43.0/24 peer=10.0.42.0/24"
264
dt and dtsh
The dt program continuously writes data to node hard drives for testing purposes; dtsh is a shell script that invokes the dt utility with special parameters.
Synopsis
dt -dir=PATH [-bs=SIZE] [-error] [{-f=NUM | -files=NUM | -p=NUM}] [-fs=SIZE] [{-gb | -mb}] [-pass=NUM] [-pat=NUM] [-ran] [-rexit] [-rpass=NUM] [-verbose] [-wpass=NUM]
Options
-bs=SIZE -dir=PATH -error -f=NUM | -files=NUM | -p=NUM -fs=SIZE -gb -mb -pass=NUM -pat=NUM -ran -rexit -rpass=NUM -verbose Size of each file block. Default is 1MB. Base directory to write files in. Required. dt will exit when an error is detected. Default value is off. Number (NUM) of files to write. Default is 1.
SIZE of each file. Default is 1GB. File size is in gigabytes. This is the default. File size is in megabytes. Number (NUM) of passes to run. Default is 1. Number (NUM) of patterns to run. Default is 14. Write random data. Only used to generate data, there is no random data read/compare. Exits after first pass when in random mode Number (NUM) of read passes to run. Default is 1. Enables progress information, which was formerly displayed by default. Default behavior is to run more silently. dt now prints a final status message at the end. Number (NUM) of write passes to run. Default is 1.
-wpass=NUM
Notes
The dt program should be copied onto every node using mapall copy dt. A dtsh script, which will launch dt on each partition, should be created next, as followis:
dtgen: !/bin/sh # Create 4 1GB files of random data ./dt -dir=/data01 -f=4 -ran -pass=1 &
265
The dtgen program should be copied onto every node using mapall copy dtgen. You can then run dt on all the nodes by using mapall --bg ../dtgen. The dt program will create a working directory QA in each of the -dir parameters specified. dt will create a logfile in each of the working directories, it looks something like:
admin@node-10-0-53-10:/data01/QA/>: cat node_10_0_53_10__data01_dt.log Wed Jan 29 19:49:40 2003 hostname=node-10-0-53-10 (node-10-0-53-10.local.avamar.com) Wed Jan 29 19:49:40 2003 -dir=/data01 Wed Jan 29 19:49:40 2003 -patterns=1 Wed Jan 29 19:49:40 2003 -pass=1 Wed Jan 29 19:49:40 2003 -f=4 Wed Jan 29 19:49:40 2003 -bs=1048576 (KB:1024 mb:1048576 gb:1073741824) Wed Jan 29 19:49:40 2003 -fs=1073741824 Wed Jan 29 19:49:40 2003 Seed = 34464415 Wed Jan 29 19:49:40 2003 pattern=0 Wed Jan 29 19:49:40 2003 filename=dt.0 Wed Jan 29 19:51:34 2003 filename=dt.0 (W:7659 G:3500) Wed Jan 29 19:51:34 2003 filename=dt.1 Wed Jan 29 19:53:32 2003 filename=dt.1 (W:7672 G:3601) Wed Jan 29 19:53:32 2003 filename=dt.2 Wed Jan 29 19:55:28 2003 filename=dt.2 (W:7697 G:3507) Wed Jan 29 19:55:28 2003 filename=dt.3 Wed Jan 29 19:57:15 2003 filename=dt.3 (W:6979 G:3269) Wed Jan 29 19:57:15 2003 58.6904 MB / Second written (1073741824 4 W:30007 G:13877) Note there are number prefixed with W:, G:, R: and C: T:
Where: W G R C T Write time in ticks. Data generation time in ticks. Read time in ticks. Compare time in ticks. Total time in ticks.
You should remove the /data0?/QA/dt.* files when dt is finished. For example: mapall --bg --all rm "/data0?/QA/dt.*"
266
dump_accounts
The dump_accounts program runs avmgr to return information from the account stripe for each specified path. The data that is output from this command is voluminous and should be redirected to a file. This file can be used as input for load_accounts utility (page 298) to reload dumped information into a dpn. This utility is primarily used as a development tool.
Synopsis
dump_accounts --hfsaddr=AVAMARSERVER --id=root -ap=PASSWORD CLIENT-PATH
Options
--ap=PASSWORD --hfsaddr=AVAMARSERVER --id=root Authenticate using PASSWORD. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Run as root.
Client Path
CLIENT-PATH Specifies a valid Avamar client path (for example, /clients/MyClient). Additional client paths can be delimited with white space.
Notes
Typically, --hfsaddr=AVAMARSERVER is set in a configuration file on the utility node. Therefore, this option is not normally required on the command line.
267
emserver.sh
The emserver.sh shell script configures the EMS.
Synopsis
emserver.sh {--changelocalap | --flush | --help | --init | --ping | --renameserver | --restart | --restore | --start | --status | --stop | --testwebapp} [--ap=PASSWORD] [--flushfile=FILENAME] [--flushtype=local] [--hfsaddr=AVAMARSERVER] [--hfsport=PORT] [--id=USER-ID] [--label=LABEL] [--labelnum=NUM] [--mchostname=HOSTNAME] [--mcport=PORT] [--mcuserap=PASSWORD] [--rootap=PASSWORD] [--update] [{--uselocalmcs | --nouselocalmcs}] [--version] [--xml]
Commands
One (and only one) of the following commands must be supplied on each command line: --changelocalap IMPORTANT: EMC internal use only. Changes the stored password for the local Avamar server. --flush --help --init --ping --renameserver Backs up (flushes) EMS data. Shows full online help (commands, options and full descriptions) for this utility, then exits. Initializes the EMS. Performs a check (ping) to determine if the EMS is listening. IMPORTANT: EMC internal use only. Changes the hostname and IP address of the authenticating Avamar server. --restart --restore --start --status --stop --testwebapp Shuts down all EMS services and restarts them. Each restart is logged. Restores EMS from a previous flush. Starts the MCS and logs the action. Returns status of each EMS service and any available performance statistics. Shuts down (stops) the EMS services. Each shut down is logged. Provides status of the Apache Jakarta Tomcat service, which is installed along with the EMS. This command added in version 3.7.
268
Options
--ap=PASSWORD --flushfile=FILENAME --flushtype=local --hfsaddr=AVAMARSERVER Specifies password for --id=USER-ID. Specifies a local flush file. Restores EMS state from a local flush. Specifies AVAMARSERVER IP address or fully qualified hostname (as defined in DNS) from which to from which to restore the EMS. Used with --init, --renameserver or --restore to directly specify which Avamar server data port to use for user authentication or restore. Supplying this option bypasses the corresponding interactive prompt. This setting is ignored if EMS authenticates using the local MCS. --id=USER-ID --label=LABEL --labelnum=NUM --mchostname=HOSTNAME Specifies an alternate user account (USER-ID) to perform restore. Specifies flush label name from which to restore the EMS. Specifies flush sequence number (NUM) from which to restore the EMS. Used with --init or --renameserver to directly specify which MCS hostname or IP address to use for user authentication. Supplying this option bypasses the corresponding interactive prompt. This setting is ignored if EMS authenticates using the local MCS. --mcport=PORT Used with --init or --renameserver to directly specify which MCS data port to use for user authentication. Supplying this option bypasses the corresponding interactive prompt. This setting is ignored if EMS authenticates using the local MCS. --mcuserap=PASSWORD Used with --init or --renameserver to directly specify the MCUser account PASSWORD on the authenticating MCS. Supplying this option bypasses the corresponding interactive prompt. This setting is ignored if EMS authenticates using the local MCS.
--hfsport=PORT
269
emserver.sh COMMAND REFERENCE --rootap=PASSWORD Used with --init or --renameserver to directly specify the root account PASSWORD on the authenticating MCS. Supplying this option bypasses the corresponding interactive prompt. This setting is ignored if EMS authenticates using the local MCS. --update --uselocalmcs Updates EMS datastore and preferences following a software upgrade. Used with --init or --renameserver to bypass the corresponding interactive prompt and specify that the EMS should authenticate locally. Shows version, then exits. Used with --status to output an XML document instead of the standard --status output.
--version --xml
Notes
This shell script added in version 3.5.0.
270
emwebapp.sh
The emwebapp.sh shell script starts and stops the Apache Jakarta Tomcat service with the JVM parameter to allow the JVM to allocate more memory.
Synopsis
emwebapp.sh {--error | --help | --restart | --run | --start | --stop | --test | --update}
Commands
One (and only one) of the following commands must be supplied on each command line: --error --help --restart --run --start --stop --test IMPORTANT: EMC internal use only. Shows online help for this utility, then exits. Performs a stop, then a start. Equivalent to running emwebapp.sh --stop, then emwebapp.sh --start. IMPORTANT: EMC internal use only. Starts the Apache Jakarta Tomcat service. Shuts down (stops) the Apache Jakarta Tomcat service. Provides status of the Apache Jakarta Tomcat service. This command added in version 3.7. --update Forces a configuration update. This command added in version 4.0.
Notes
The emwebapp.sh shell script must be run as root. This shell script added in version 3.5.0.
271
errchk
The errchk program that scans a series of log files created by the operating system and Avamar server software and creates a time stamped plain text log file, which can be sent to EMC Technical Support for examination. The errchk program does not attempt to assess the server health. It is merely a filtering tool. Therefore, the generated log file must be examined by EMC Technical Support to determine if there are issues that require correction or additional investigation. No user data that has been backed up from client systems is included in the errchk output file.
Output
The errchk program generates a text file with a time/date stamp in the current working directory. The output log filename has the format errchk-YYYY-MM-DDHHMM.txt (errchk-2003-03-07-0958.txt). Each errchk log file can potentially contain the following kinds of entries: Operating system kernel and I/O errors for all nodes retrieved from /var/log/messages files Internal Avamar server errors retrieved from gsan.log files Errors from cron jobs run on utility node Check of node state, access mode and loadavg on all storage nodes All Avamar server stripes that are not ONLINE All checkpoints and whether they have been hfschecked Full NODELIST data for each server node
Notes
What to Look For Any kernel and I/O errors retrieved from /var/log/messages files are probably
reason for concern. In particular, hard drive or DMA drive errors probably indicate hardware trouble.
Any Avamar server errors retrieved from gsan.log files should be reported to EMC Technical Support. While not all errors are serious, they should be reviewed by EMC Technical Support. Environmental conditions such as temporary network outages can generate server errors that are temporary and have no lasting impact after the situation is corrected. Scrutinize any errors from cron jobs run on utility node. This typically indicates that a regular maintenance process (checkpoint, daily data integrity check and garbage collection) might not have run correctly. These should be reported to EMC Technical Support. Scrutinize the check of node state, access mode and loadavg on all storage nodes. Verify that all nodes are ONLINE, all access modes are normal and all loadavgs are 2 or less. If any nodes are OFFLINE or the access mode is anything other than normal contact EMC Technical Support. Scrutinize the list of all Avamar server stripes that are not ONLINE. If any stripes are designated OFFLINE or OFFLINE_MEDIA_ERROR, contact EMC Technical Support. If stripes are MIGRATING, then a node data migration operation is being performed. AVAMAR 4.1 TECHNICAL ADDENDUM 272
errchk COMMAND REFERENCE Scrutinize the list of checkpoints and verify that the last checkpoint was created within the last 12 hours and that at least one checkpoint passed a daily data integrity check within the last 24 hours.
Example
errchk
Creating errchk-2003-03-19-1207.txt... Searching /var/log/messages files Searching GSAN log files Searching MC log files Searching cron job log files Checking node status Checking stripe status Getting list of checkpoints Getting full nodelist Done: 13492 errchk-2003-03-19-1207.txt
273
evening_cron_run
The evening_cron_run program runs the evening cron job. It invokes cp_cron (page 224) and gc_cron (page 278) and is itself intended to be invoked by cron_env_wrapper (page 228). To run this script directly from a command shell, enter the following: cron_env_wrapper evening_cron_run IMPORTANT: This program must be run as user dpn. This program attempts to load the dpnid OpenSSH key from the dpn user .ssh directory. This program will fail if it is run as a different user.
274
expire-snapups
The expire-snapups program writes a script to stdout, which when run, changes backup expiration dates.
Synopsis
expire-snapups [--after=DATE] [--before=DATE] [--days=NUM] [--domain=CLIENT-DOMAIN] [--help] [--include_mc_backups] [CLIENT-PATH ...]
Options
--after=DATE If supplied, all backups created after this date are eligible to be expired. Default DATE setting is 90 days ago. If supplied, all backups created before this date are eligible to be expired. Default DATE setting is now (the time at which expire-snapups is invoked). Specifies backup expiration date and time as the number (NUM) of whole days from today. Default value is 90 days from the initial backup creation date. If supplied, all clients within the specified CLIENTDOMAIN are processed. Shows help, then exits. If supplied, all clients within the special MC_BACKUPS domain are also processed. Because MC_BACKUPS is a special system domain, the default behavior is to not process these clients unless this option is explicitly supplied.
--before=DATE
--days=NUM
Client Path
CLIENT-PATH Specifies a valid Avamar client path (for example, /clients/MyClient). Additional client paths can be delimited with white space. IMPORTANT: Client paths must always terminate with a valid Avamar client name. You cannot specify a client domain in this manner. Use --domain when you want to process all clients within a particular domain. If not supplied, all clients are processed. This is the default setting.
275
Notes
The DATE specifier can be any date string acceptable to date(1). For GNU date(1), which on Linux is /bin/date, DATE can be just about any common date string, including such phrases as 2 weeks ago. This program added in version 3.0.
276
gathergsankeydata
The gathergsankeydata program gathers essential server information and writes it to an output file, which is then used to generate a server license key file. The gathergsankeydata program can be run interactively or non-interactively. Non-interactive mode is primarily used to support use of this utility by other utilities (for example, avw_install). Refer to the Avamar Server Software Installation Manual for more information about avw_install.
Synopsis
gathergsankeydata [--account_id=CUSTOMER-ID] [--asset_reference_id=ASSET-ID] [--interactive] [--output=PATH] [--run]
Options
--account_id=CUSTOMER-ID Specifies numeric CUSTOMER-ID, which is found on each EMC Purchase Order Confirmation. Specifies system asset reference ID, which is found on each EMC Purchase Order Confirmation. Specifies whether to run gathergsankeydata interactively (default) or non-interactively. Directory path and filename of output FILE. Default value is gsankeydata.xml). Use dash (-) for stdout. Specifies run (default) or dry-run modes.
--asset_reference_id=ASSET-ID
--interactive
--output=FILE
--run
Notes
Files used: TOP/var/probe.out TOP/var/mail.log Static node information. Email log file.
277
gc_cron
Garbage collection cron job. The gc_cron program is intended to be run as a cron job that will perform garbage collection. It is invoked from morning_cron_run (page 323) and evening_cron_run (page 274).
Synopsis
gc_cron [--convcount=NUM] [--conversion ] [--convtime=SEC] [--full] [--gccount=NUM] [--killsnapups=SEC] [--limitadjust={+ | -}LIMIT] [--passes=NUM] [--pollrate=SEC] [--refcheck] [--throttlelevel=PERCENT] [--timeout=SEC] [--usehistory]
Options
--convcount=NUM --conversion --convtime=SEC Specifies maximum number (NUM) of stripes that will be converted on each node following this gc_cron session. Specifies whether or not convert pre-3.5 data stripes into version 3.5 data stripes. Default value is true. Specifies maximum amount of time that avmaint conversion will be allowed to run following each gc_cron session. Specifies the kind of status to be obtained and displayed. Refer to avmaint gcstatus (page 95) for additional information. This is not the default setting. Specifies number (NUM) of index stripes per node to garbage collect. Default value is 16. If supplied, any backups in progress are forcibly stopped (killed) after waiting the specified number of seconds (SEC). If not supplied or set to zero, backups are not forcibly stopped (killed). This option added in version 4.0. --limitadjust= {+ | -}LIMIT Used with --usehistory to increase or decrease computed garbage collect quota by LIMIT percent. Default setting is zero (0). For example, if history computes that NN GB should be garbage collected, then using --limitadjust=+50 increases that amount to NN + (NN*50%) GB. Alternatively, supplying --limitadjust=-25 decreases that amount to NN - (NN*25%) GB. This option added in version 4.1. --passes=NUM Specifies maximum number (NUM) of garbage collection passes. Default value is 100 passes.
--full
--gccount=NUM --killsnapups=SEC
278
gc_cron COMMAND REFERENCE --pollrate=SEC Specifies interval in seconds (SEC) at which the server is polled for garbage collection status. Default setting is 60 seconds. Forces check of composite references during garbage collect or HFS check, respectively. Specifies throttling percentage (PERCENT). This value reduces the garbage collect operations system resource (CPU, disk, bandwidth, and so forth) usage by this percentage. Default setting is zero (that is, no throttling), which means that the garbage collect operations system resource usage is potentially 100%). This option added in version 4.0. --timeout=SEC Specifies maximum number of seconds (SEC) garbage collect process will be allowed to run. Default value is 3600 seconds (1 hour). Enables or disables garbage collect history feature, which limits the amount of garbage collection performed to an amount that will free up at least as much space as has been used by the daily average of all backups over the last several days. Default is true. This option added in version 4.1.
--refcheck --throttlelevel=PERCENT
--usehistory
Notes
IMPORTANT: gc_cron must be run as user dpn. Beginning with version 4.0, if gc_cron is forcibly terminated by an external signal (for example, using the kill command), it will stop garbage collection on the storage server (also known as GSAN). Beginning with version 3.5, gc_cron initiates a stripe conversion session by invoking avmaint conversion (page 86) on exit and passing in the --convcount and --convtime options.
279
gen-ssl-cert
The gen-ssl-certprogram installs a temporary Apache web server 1.1 SSL cert and restarts the web server.
Synopsis
gen-ssl-cert [--debug] [--help] [--verbose]
Options
--debug --help --verbose Prints utility session information but does not actually perform the actions. Shows help, then exits. Provides maximum information.
Notes
The gen-ssl-certprogram must be run as root. Original files are backed up and saved as: /etc/httpd/conf/ssl.crt/server.crt.orig /etc/httpd/conf/ssl.key/server.key.orig
280
gethostbyname
The gethostbyname program shows name resolution in the same way that programs using gethostbyname() and gethostbyaddr() libc functions do.
Synopsis
gethostbyname HOST-NAME
Notes
This utility is primarily intended for use by EMC personnel only.
Examples
This command returns all hostnames resolving to ntp. gethostbyname ntp
gethostbyname(ntp) => 192.168.100.3, 192.168.100.5, 192.168.100.7 gethostbyaddr(192.168.100.3) => vortex.example.com gethostbyaddr(192.168.100.5) => central.example.com gethostbyaddr(192.168.100.7) => epicenter.example.com
281
getlogs
The getlogs program gathers important log files from all server nodes and writes them to local utility node directories where they can be viewed and analyzed to support various maintenance and troubleshooting activities. The precise mechanism for accomplishing this is: 1. The getlogs program copies getnodelogs (page 283) to each node in the system and runs it. 2. The getnodelogs program gathers the important log files from that particular node and compresses them into a single tar file (nodelogs.tgz). 3. The getlogs program creates a master tar file on the utility node, which contains the individual nodelogs.tgz. The getlogs program accepts an optional filename (FILE). If not supplied, the default filename logs.DATE.tar (page 444) is used, where DATE is an eight character date code (YYYYMMDD) and six-character timestamp (HHMMSS).
Synopsis
getlogs [--server={today | yesterday | week | restart | NUM-DAYS}] [--verbose] [FILE]
Options
--server={today | yesterday | week | restart | NUM-DAYS} Controls how much of the gsan logs should be downloaded. One of the following: today yesterday week restart NUM-DAYS Download the previous 24 hours worth of server log files. Download the previous 48 hours worth of server log files. Download the previous 168 hours worth of server log files. Download all gsan log files since the last restart. Download this number of days (NUM-DAYS) worth of server log files.
If --server is not supplied, all server log files are downloaded. This option added in version 2.0. --verbose Provides maximum information. This option added in version 2.0.
282
getnodelogs
The getnodelogs program is invoked by getlogs (page 282), which copies it to each node in the system. When run locally from each node, getnodelogs gathers the important log files from that particular node and compresses them into a single tar file (nodelogs.tgz).
283
getsnapupstats
The getsnapupstats program invokes avmgr and avtar to return a list of backups that were successfully performed on the specified host, then downloads the backup statistics from the server into a local working directory. The script then creates a compressed tar file (.tgz) that contains all of the statistics. This binary file can then be sent to EMC Technical Support for examination. No user data that has been backed up from client systems is included in getsnapupstats data. Data collected for each backup includes: archive_info filestats groups machine.xml mbr0.bin mounts partitiontables.xml smartdata.xml statsfile users vbr0-0.bin volumes.xml workorder Original avtar command line. Statistics on the top 500 files. List of group IDs. Client machine characteristics (Windows only). Master boot record (Windows only). List of mount points. Partition table (Windows only). SMART hard drive data (Windows only). Progress reports sent to server every 30 seconds. List of user IDs. Volume boot record (Windows only). List of volumes (Windows only). Original workorder provided by MCS.
Synopsis
getsnapupstats --path=LOCATION [--id=USER@AUTH] [{--ap=PASSWORD | --password=PASSWORD}] [OPTIONS] [--max=n]
Options
--path=LOCATION Specifies a hierarchical LOCATION on the Avamar server. This option is relative to your current home location, unless a slash (/) is used as a prefix to the path designation, in which case an absolute path is assumed.
284
getsnapupstats COMMAND REFERENCE --id=USER@AUTH Authenticate as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --ap=PASSWORD | --password=PASSWORD OPTIONS PASSWORD for the --id=USER@AUTH account. avmgr (page 144) or avtar (page 171) options. These options are passed directly to avmgr and avtar without interpretation. Optional flag to specify the maximum number of backups to examine. Backups are examined in most recent to least recent order. For example, supplying --max=10 returns statistics for the ten most recent backups. Default behavior is to return all backup statistics for the HOST.
--max=n
Output
The getsnapupstats program creates a local directory based on the client hostname supplied with --path, then creates a compressed tar file (.tgz) containing everything in the local directory. You should delete this working directory after running the utility.
Example
getsnapupstats --path=/clients/my-client --id=root --ap=*****
Restoring my-client/10:avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=10 --target=my-client/10 Restoring my-client/9: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=9 --target=my-client/9 Restoring my-client/8: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=8 --target=my-client/8 Restoring my-client/7: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=7 --target=my-client/7 Restoring my-client/6: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=6 --target=my-client/6 Restoring my-client/5: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=5 --target=my-client/5 Restoring my-client/4: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=4 --target=my-client/4 Restoring my-client/3: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=3 --target=my-client/3 Restoring my-client/2: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=2 --target=my-client/2 Restoring my-client/1: avtar -x --quiet --keep-old=false --path=/clients/my-client --id=root --ap=***** --internal .system_info --labelnum=1 --target=my-client/1 Creating my-client.tgz:tar cfz my-client.tgz my-client my-client.tgz created. You may delete the local temporary directory 'my-client'.
285
health_check.pl
The health_check.pl Perl script returns a status report for the Avamar server. The health_check.pl Perl script displays a summary that contains status for the following: Subsystem summary Per-node GSAN Checkpoint Checkpoint validation (hfscheck) Garbage collection Key configuration parameters Per-node storage capacity Per-node time server Starting with Avamar 4.1 the dpnugprep package contains a new version of health_check.pl. This version of health_check.pl runs either directly from the command line or as a subprogram of avupos. NOTE: The avupos program automates the OS upgrade from RHEL3 (32-bit version) to RHEL4 (64-bit version) on nodes running Avamar release 3.x. The EMC Avamar 4.1 System Upgrade Manaul provide more information about avupos The health_check.pl Perl script requires login to the admin user in an ssh-agent context with either the dpnid or admin_key loaded.
Synopsis
su - admin ssh-agent bash ssh-add ~/.ssh/dpnid health_check.pl [--help] [--log=PATH] [--norequire_sched_up] [--norequire_unattended_startup]
Options
--help --log=PATH --norequire_sched_up Shows help, then exits. Specifies a logfile PATH. Bypasses the check to determine if the scheduler is running. The default setting is to check the running status of the scheduler.
286
health_check.pl COMMAND REFERENCE --norequire_unattended_startup Bypasses the check to determine if unattended startup is enabled on a single-node server. The default setting is to check the status of unattended startup on a single-node server.
Notes
The health_check.pl Perl script is located in /usr/local/avamar/bin/. Run health_check.pl before and after performing a manual upgrade. This Perl script added in version 4.0.1.
Output
The following excerpt is an example of the output from the health_check.pl program:
============================================== ==== Health Check Summary ==== ============================================== Logfile: /tmp/health.20080312-1405 Total Number of Errors: 3 Errors to be reviewed: 20080312-1405.49: ERROR: Last hfscheck was not successful! (errcode=4004) 20080312-1405.49: ERROR: Garbage collecting less than 10GB per pass! (averaging 3.79 GB) 20080312-1405.51: ERROR: Error(s) found in err.log file. Please review /tmp/health.20080312-1405 for more info.
The Health Check Summary also includes status for the dpnctl program. The following output illustrates status for dpnctl:
dpnctl: dpnctl: dpnctl: dpnctl: dpnctl: dpnctl: INFO: INFO: INFO: INFO: INFO: INFO: gsan status: ready MCS status: up. EMS status: up. Scheduler status: up. Maintenance operations status: enabled. Unattended startup status: disabled.
287
hfscheck_cron
The hfscheck_cron program is primarily intended to be run as a cron job in order to start an HFS check. In this capacity, it is typically called from morning_cron_run (page 323). However, hfscheck_cron can also be run directly to perform an on-demand HFS check or create a list of valid checkpoints in the system. Refer to Checkpoint Validation (HFS Checks) (page 15) for additional detailed technical information about the HFS check feature.
Synopsis
hfscheck_cron [{--checks=DESCRIPTOR | --full | --metadata | --parity | --refcheck}] [--checknode=MODULE.NODE] [--checkpoint=CP-ID] [--checkpoints=FILE] [--cleanup] [--concurrentdatastripes=NUM] [--concurrentindexstripes=NUM] [--concurrentparitystripes=NUM] [--deadline=MIN] [--gccount=NUM] [--keepmin] [--oldstyle] [--scheduleconffile=FILE] [--suspend] [--throttlelevel=NUM] [--useschedule] [--xmlperline=NUM]
Options
--checks=DESCRIPTOR DESCRIPTOR must be a valid enabling descriptor string that describes which checks are to be performed on which specific stripe classes. Stripe classes can be any or all of the following: h p u HFS stripes. Persistent stripes. Accounting stripes.
Checks can be any or all of the following: a c p r Atomic data sweeps. Composite data sweeps. Parity checks. Reference checking.
For example, this descriptor: hpu+acpr Is equivalent to the default behavior (perform all checks on all stripe classes).
288
hfscheck_cron COMMAND REFERENCE --checks=DESCRIPTOR (continued) You can also constrain an HFS check to only process a percentage of the total stripes by prefixing a desired processing percentage (as an integer between 1 and 99). For example: 50:hpu Default value is 0 (process 100% of eligible stripes). Refer to HFS Check Enabling Descriptor (page 17) for additional information about enabling descriptor syntax. This option added in version 3.5. --checknode=MODULE.NODE Enables single-node HFS check (page 21) on this specific MODULE.NODE. This option added in version 3.6. --checkpoint=CP-ID Specifies which checkpoint should be processed. If supplied, HFS check is performed on this checkpoint CP-ID. If not supplied, HFS check is performed on the most recent unvalidated checkpoint file. This option added in version 3.5. --checkpoints=FILE Specifies the location and filename for the checkpoints listing file. Default location is /usr/local/avamar/var/checkpoints.xml. If supplied, all temporary working directories and files are removed when the HFS check completes. This is the default setting. Supplying --nocleanup preserves these directories and can be used to recover the HFS check log files. --concurrentdatastripes=NUM Specifies maximum number of data stripes that are simultaneously processed on a single node during a index sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5. --concurrentindexstripes=NUM Specifies maximum number of index stripes that are simultaneously processed on a single node during an index sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5.
--cleanup
289
hfscheck_cron COMMAND REFERENCE --concurrentparitystripes=NUM Specifies maximum number of parity stripes that are simultaneously processed on a single node during a parity sweep. The higher the value, the more concurrent work is being performed. Default value is 1. This option added in version 3.5. --deadline=MIN If supplied, specifies HFS check deadline (page 21) in minutes (MIN). Default value is zero (no deadline). This option added in version 3.6. --full Overrides the default schedule for the day in order to perform a full HFS check. Functionally equivalent to --checks=0:hpu+acpr. IMPORTANT: Cannot be combined with --checks. This option added in version 4.0. --gccount=NUM Specifies maximum number of index stripes per node that are to be reference checked. Default value is 16. This option added in version 3.5. --keepmin Enables the keep minimal checkpoints feature for this session. Refer to The Keep Minimal Checkpoints Feature (page 13) for additional information. This option added in version 4.1. --metadata Overrides the default schedule for the day in order to perform a metadata-only HFS check. Functionally equivalent to --checks=0:hpu+cp. IMPORTANT: Cannot be combined with --checks. This option added in version 4.0. --oldstyle Reverts the command behavior to earlier versions. In particular, only info codes 4003 (success) and 4004 (failure) are used. The new behavior reserves code 4003 for a full successful check, code 4004 for any observed failure and code 4006 for successful partial, single or reduced checks. This option also affects the display of the terminating cplist command (page 225). This option added in version 3.7.1.
290
hfscheck_cron COMMAND REFERENCE --parity Overrides the default schedule for the day in order to perform a parity-only HFS check. Functionally equivalent to --checks=0:hpu+p. IMPORTANT: Cannot be combined with --checks. This option added in version 4.0. --refcheck Overrides the default schedule for the day in order to perform a refcheck-only HFS check. This option was deprecated for versions 3.5 through 4.0 in favor of --checks. Beginning with version 4.0, it is again valid and functionally equivalent to --checks=100:hpu+r. IMPORTANT: Cannot be combined with --checks. --scheduleconffile=FILE Use the specified configuration FILE to schedule hfscheck_cron operations. The default hfscheck_cron schedule file is /usr/local/avamar/var/hfsscheduleconf.xml. This option added in version 4.0. --suspend Specifies that dispatchers should be suspended during HFS check program start up (at the same time that server access mode is set to read-only). This is not the default setting. This option added in version 3.6. --throttlelevel=NUM Specifies the amount of throttled resources (measured as a percentage of CPU capacity) that can be allocated to other tasks. For example, a setting of 20 begins throttling once CPU usage exceeds 80%. Values in excess of 100 are treated as 100% (maximum throttling); a value of zero specifies that no throttling is to be performed. If a deadline is specified, the throttle level indicates a maximum allowable throttle level, and the throttle level starts at half this value. Default value is 20 if no deadline is specified; 40 if deadline is specified. Refer to HFS Check Throttling (page 19) for additional information about throttle levels and deadlines. This option added in version 3.5.
291
hfscheck_cron COMMAND REFERENCE --useschedule Allows for the use of the schedule configuration file with HFS check. Default setting is false. IMPORTANT: This option will be ignored if --checks, --full, --metadata, --parity or --refcheck is also supplied. This option added in version 4.0 --xmlperline=NUM Controls number (NUM) of XML attributes per line.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged: --forcepick=CP-ID IMPORTANT: Deprecated in version 3.5. Performs an HFS check on this (CP-ID) checkpoint. --maingsandown IMPORTANT: Deprecated in version 3.5. Causes the HFS check to refrain from any attempt to communicate with the main server. IMPORTANT: --forcepick=CP-ID must also be supplied. --paritymode=MODE IMPORTANT: Deprecated in version 3.5 in favor of --checks. Specifies how parity stripes should be checked during an hfscheck. MODE must be one of the following: columnxor none Default method of parity checking is performed. No parity checking is performed.
Default value is columnxor. This option added in version 3.5. --readonly IMPORTANT: Deprecated in version 3.5. Forces server into a read-only state long enough for hfscheck_cron to successfully start on all storage nodes. This is the default setting. This option added in version 3.0.
Notes
The hfscheck_cron program must be run as user dpn. If a schedule configuration file is not present as specified by the --scheduleconffile=FILE option or the file contains errors, hfscheck_cron will run a full HFS check.
292
Input File
You can only specify one checktype for any given day of the week. If an entry is missing for one or more days, then hfscheck_cron will run a full HFS check. Multiple checktypes for the same day of the week are considered errors. This will invalidate the schedule configuration file and halt HFS checks until a valid schedule configuration file is present as specified by the --scheduleconffile=FILE option. However, hfscheck_cron will still run a full HFS check when it is invoked on the command line even if the schedule configuration file is invalid. The schedule configuration file must conform to the following format:
<hfscheckactivitylist> <hfscheckactivity checktype="full" dayofweek="Saturday" extraflags="--throttlelevel=0"/> <hfscheckactivity checktype="metadata" dayofweek="Monday"/> <hfscheckactivity checktype="parity" dayofweek="Tuesday" extraflags="--throttlelevel=20"/> <hfscheckactivity checktype="refcheck" dayofweek="Wednesday"/> <hfscheckactivity checktype="50:hp+ac" dayofweek="Thursday"/> <hfscheckactivity checktype="75:hpu+acpr" dayofweek="Friday"/> </hfscheckactivitylist>
293
hfscheck_kill
The hfscheck_kill program gracefully terminates (kills) a currently running hfscheck process.
hfsclean
The hfsclean program deletes all data currently stored in the Avamar server. It is used by the start.dpn (page 382) and start.nodes scripts. IMPORTANT: Documentation for this script is provided for reference purposes only. Do not run hfsclean directly from the command line.
hfssetup
The hfssetup program creates the /usr/local/avamar/etc/avamar.cfg file. This file is a configuration file for the web services. IMPORTANT: Documentation for this script is provided for reference purposes only. Do not run hfssetup directly from the command line.
initacnt
The initacnt program is run once during Avamar server initialization. It creates base administrative user accounts and default directories in the Avamar server.
Synopsis
initacnt --hfsaddr=AVAMARSERVER --password=PASSWORD
Options
--hfsaddr=AVAMARSERVER --password=PASSWORD AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Internal Avamar root user account PASSWORD (not the operating system root password).
294
java_update.sh
The java_update.sh shell script updates the specific Java Runtime Environment (JRE) version used by various features and functions.
Synopsis
java_update.sh {client | ems | mccli | mds | server} [JRE-DIR]
Commands
client ems mccli mds server Update JRE used by the Administrator client scripts. Update JRE used by Avamar Enterprise Manager server (EMS). Update JRE used by Avamar Management Console Command Line Interface (MCCLI). Update JRE used by the Metadata Search (MDS) feature. Update JRE used by the MCS scripts.
Options
JRE-DIR JRE root directory (for example, /usr/java/jre1.5.0_12). If not supplied, java_update.sh will search for the latest installed JRE and provide that root directory as the default value.
295
lm COMMAND REFERENCE
lm
Avamar login manager. This process provides access to external authentication databases, which allows Avamar to use pre-existing login user name/password information for Avamar authentication. Without the login manager, Avamar can only use its internal authentication mechanism. When requests from an Avamar client specify the use of an external authentication domain, Avamar redirects login messages to the external domain by way of the lm processes. The external domain returns status to the login manager. An lm process is expected to be running on each utility node. The login manager requires multiple configuration files in order to operate properly. IMPORTANT: The login manager must be restarted after any change to the configuration files.
Synopsis
lm [--debug] [--encrypt={ssl | tcp}] [--encrypt-strength={cleartext | high | medium}] [--flagfile=FILE] [--help] [--helpx] [--helpxml] [--memman=NAME] [--memmantrigger=NAME] [--port=PORT] [--sysdir=PATH] [--timeout=SEC] [--uflagsdebug] [--usage] [--verbose] [--version]
Options
--debug Prevents login manager from entering background mode and automatically enables verbose messaging. Sets encryption type to one of the following: ssl tcp --encrypt-strength= {cleartext | medium | high} Use Secure Sockets Layer encryption. Do not use SSL. This is the default setting.
--encrypt={ssl | tcp}
Sets encryption strength to one of the following: cleartext high medium Corresponds to the NULL-SHA cipher. Corresponds to the AES256-SHA cipher. This is the default setting. Corresponds to the AES128-SHA cipher.
296
lm COMMAND REFERENCE --flagfile=FILE Specifies FILE containing a list of options and values that will be processed by this utility as if they were included on the command line. Default is /usr/local/avamar/etc/usersettings.cfg. Shows full online help (commands, options and full descriptions) for this utility, then exits. Shows full online help for this utility, including extended flags, then exits. Shows full online help for this utility in xml format, then exits. IMPORTANT: EMC internal use only. Memory debugging MODE. --memmantrigger=MODE IMPORTANT: EMC internal use only. Extra memory debugging MODE. --port=PORT Specifies an alternate data PORT for login manager communication. Default value is 700 (which requires operating system root privileges). PATH to the directory containing Avamar server files. Default value is /etc/avamar. Time limit in seconds (SEC) applied to each login attempt. Default value is 10 seconds. Specifies flag parsing debug mode. Shows abbreviated online help (commands and options only, no descriptions) for this utility, then exits. Provides maximum information. Shows version, then exits.
--verbose --version
Notes
The lm process is typically activated during system startup. The process will automatically run in the background unless debug mode is asserted.
297
load_accounts
The load_accounts program reads one or more files previously generated by dump_accounts (page 267), then creates the clients, domains, accounts and users described in the file on the specified Avamar server. This program is primarily used as a development tool.
Synopsis
load_accounts [-ap=PASSWORD] [--debug] [--hfsaddr=AVAMARSERVER] [--id=root] [--showonly] FILE1 [, FILE2, ... ]
Options
--ap=PASSWORD --debug -hfsaddr=AVAMARSERVER --id=root --showonly FILE1 [, FILE2, ... ] Authenticate using PASSWORD. Enable debug mode. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Run as root. Show only, do not run. One or more files previously generated by dump_accounts (page 267). Separate multiple entries with a comma (,).
Notes
Typically, --hfsaddr=AVAMARSERVER is set in a configuration file on the utility node. Therefore, this option is not normally required on the command line.
logmrg
The logmrg program can only be run after getlogs (page 282). It parses gsan.log files in MODULE.NODE subdirectories created by getlogs and creates a unified log file that is in time order. The output goes to stdout and should therefore be redirected to a file.
298
mapall
The mapall program runs the same command on all of the nodes in the Avamar server. It uses probe.out (page 477) to resolve MODULE.NODE designations into IP addresses.
Synopsis
mapall [--all] [--bg] [--capture] [copy FILE] [debug] [get FILE] [--nodes=NODELIST] [--noerror] [--noid] [--parallel] [--probedir=PATH] [--quiet] [redir] [--user=USER-ID][--verbose] COMMAND [; COMMAND2 ; ...] COMMAND can be any valid operating system command on the remote system. Commands that contain wildcards or pipes (*,?,|) must be enclosed within single quotes ('*'). Multiple commands can be entered if separated by semicolons. Single quotes must be used in this case as well.
Options
--all Runs COMMAND on all nodes (utility and storage) in the system. If not supplied, COMMAND is only run on the storage nodes. Runs COMMAND in the background and does not wait for it to complete. If --capture is not supplied, COMMAND output is interleaved. Copies FILE to each of the target nodes. This is done by packing up the specified file(s) in a GNU gzipped tar archive, shipping the tar archive to the target nodes and unpacking them. IMPORTANT: On solaris, this command requires that GNU tar be installed in /usr/local/bin. debug get FILE Turns on extended debugging information (for example, which nodes map to which IP addresses). Copies FILE from each of the target nodes. The file from each node is deposited into a directory named for the target node. For example, mapall get /etc/hosts places a copy of each nodes hosts file into 0.0/hosts, 0.1/ hosts, 0.2/hosts, and so forth directories. Runs COMMAND on these nodes. The following shortcuts are supported: #.# all storage nodes #.s all utility nodes #.m administrator node (deprecated) For example: 1.# is all storage nodes on module-1, 0.2,1.5 is nodes 0.2 and 1.5, #.s,#.m,#.# is all utility and storage nodes. Default value is #.#.
--nodes=NODELIST
299
mapall COMMAND REFERENCE --noerror --noid --parallel --probedir=PATH --quiet redir Continue if error occurs. Default behavior is to terminate script session after an error occurs. Does not print node ID before the COMMAND is run on each node. Runs COMMAND on each node at the same time. Specifies a PATH to a directory containing a valid probe.out file (page 477). Turns off all messages (status and warnings). Causes the standard output of COMMAND to be captured and retrieved into a subdirectory named for the target node. For example, mapall redir ls /etc will place a copy of the output into 0.0/ls.out, 0.1/ls.out, and so forth. Supplying this options overrides operating system USERID stored by the SYSPROBEUSER environment variable (that is, run mapall as this USER-ID). Typically, --user=USER-ID is used to specify that mapall be run as root, dpn or admin users. NOTE: Use of --user requires that the correct authorization exist. In particular, in order to perform mapall commands as the root or admin user, you must load the dpnid private key into your ssh-agent context. The admin_key only allows for performing mapall commands as the admin user. --verbose Provides maximum information.
--user=USER-ID
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option. This utility reads a default operating system user ID from the SYSPROBEUSER environment variable (page 425). If SYSPROBEUSER is not set, then remote commands are run as user admin.
Examples
This command copies the files specified from the specified nodes to the local machine within directories named for the source node: mapall get gsan.out
300
mapall COMMAND REFERENCE This command uploads the file gsan.out from every storage node and stores it on the local machine in a group of files named NODE/gsan.out (for example, 0.0/ gsan.out and 0.1/gsan.out, and so forth): mapall [options] copy FILE ... This command copies myscript from the local machine to every storage node: mapall copy myscript This command runs the Linux date command on every node: mapall --all date This command returns the number of storage (gsan) processes running on each node: mapall --noerror ps auxww | grep gsan | grep -v grep | wc -l
301
mcdbmaint.sh
IMPORTANT: Beginning with version 1.2.1, this utility has been deprecated in favor of dbmaint.sh (page 231). The mcdbmaint.sh shell script performs maintenance on the MCS PostgreSQL database (mcdb). It performs the same functions as mcdbsql.sh (page 303). The only difference is that mcdbmaint.sh requires the MCS to be stopped before it can be run; mcdbsql.sh can be run when the MCS is running.
Synopsis
mcdbmaint.sh {db_1.000_to_1.001.sql | db_count.sql | dbcreate.sql | dbmaint.sql | ec_update.sql | show_version s.sql}
Options
mcdbmaint.sh accepts one (and only one) of the following .sgl files as a command-line option: db_1.000_to_1.001.sql Updates the mcdb schema from version 1.000 to version 1.001. NOTE: Other similar files might also be present in /usr/local/avamar/bin. These files would perform a similar upgrade from or to a different mcdb schema version. db_count.sql dbcreate.sql dbmaint.sql ec_update.sql show_version s.sql Returns total number of records in all mcdb tables. Creates the mcdb tables. Frees unused hard drive space by running an SQL vacuum command against the mcdb. Forces an upgrade of the event catalog by resetting the version numbers. Returns schema version.
Notes
IMPORTANT: mcdbmaint.sh script can only be invoked when the MCS is stopped. When invoked, mcdbmaint.sh performs the following: 1. Stops the PostgreSQL (if it is running) to ensure a consistent database. 2. Starts the database. 3. Runs the PostgreSQL commands in the specified file. AVAMAR 4.1 TECHNICAL ADDENDUM 302
mcdbsql.sh
IMPORTANT: Beginning with version 1.2, this utility has been deprecated in favor of dbmaint.sh (page 231). The mcdbsql.sh shell script performs maintenance on the MCS PostgreSQL database (mcdb). It performs the same functions as mcdbmaint.sh (page 302). The only difference is that mcdbmaint.sh requires the MCS to be stopped before it can be run; mcdbsql.sh can be run when the MCS is running.
Synopsis
mcdbsql.sh {db_1.000_to_1.001.sql | db_count.sql | dbcreate.sql | dbmaint.sql | ec_update.sql | show_version s.sql}
Options
mcdbsql.sh accepts one (and only one) of the following .sgl files as a commandline option: db_1.000_to_1.001.sql Updates the mcdb schema from version 1.000 to version 1.001. NOTE: Other similar files might also be present in /usr/local/avamar/bin. These files would perform a similar upgrade from or to a different mcdb schema version. db_count.sql dbcreate.sql dbmaint.sql ec_update.sql show_version s.sql Returns total number of records in all mcdb tables. Creates the mcdb tables. Frees unused hard drive space by running an SQL vacuum command against the mcdb. Forces an upgrade of the event catalog by resetting the version numbers. Returns schema version.
Notes
When invoked, mcdbsql.sh performs the following: 1. Stops the PostgreSQL (if it is running) to ensure a consistent database. 2. Starts the database. 3. Runs the PostgreSQL commands in the specified file.
303
mcfeature
The mcfeature program reports which optional features are installed, configured or running on the MCS.
Synopsis
mcfeature FEATURE CHECK
Features
mds Metadata search.
Checks
installed configured running
Notes
This program added in version 4.1.
304
mcgui.bat
The mcgui.bat DOS batch file launches the Avamar Administrator graphical management console on Windows platforms. Any options supplied on the command line will populate graphical Login dialog box fields. Default location for this script is the C:\Program Files\avs\bin folder.
Synopsis
mcgui.bat [--domain DOMAIN] [--mcsaddr AVAMARSERVER] [--mcspasswd PASSWORD] [--mcsuserid USER-ID]
Options
--domain DOMAIN --mcsaddr AVAMARSERVER --mcspasswd PASSWORD --mcsuserid USER-ID Specifies top-level client domain for this Avamar Administrator session. Default value is root (/). Specifies AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Specifies PASSWORD for the --mcsuserid account. Specifies Avamar administrative user account (USER-ID) that will be used to launch the Avamar Administrator graphical management console.
305
mcgui.sh
The mcgui.sh shell script launches the Avamar Administrator graphical management console on Linux platforms. Any options supplied on the command line will populate graphical Login dialog box fields. Default location for this script is the client /usr/local/avamar/bin directory.
Synopsis
mcgui.sh [--domain DOMAIN] [--mcsaddr AVAMARSERVER] [--mcspasswd PASSWORD] [--mcsuserid USER-ID]
Options
--domain DOMAIN --mcsaddr AVAMARSERVER --mcspasswd PASSWORD --mcsuserid USER-ID Specifies top-level client domain for this Avamar Administrator session. Default value is root (/). Specifies AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Specifies PASSWORD for the --mcsuserid account. Specifies Avamar administrative user account (USER-ID) that will be used to launch the Avamar Administrator graphical management console.
306
mcgui_login.bat
The mcgui_login.bat DOS batch file launches the Avamar Administrator graphical management console from a Windows command line, a script or from within another Windows application.
Synopsis
mcgui_login.bat JRE-DIR MCS-USER-ID MCS-PSSWD DOMAIN MCS-ADDR
Options
IMPORTANT: All of the following command line options must be supplied in the exact order shown in the synopsis and as follows: Specifies location of the Sun Java installation. Specifies Avamar administrative user account (MCS-USER-ID) that will be used to log into Avamar Administrator. Specifies MCS-USER-ID password. Specifies MCS-USER-ID account domain. This should be root (/) unless you are managing sub-domains. Refer to your Avamar System Administration Manual for additional information. Specifies MCS network name or IP address.
MCS-ADDR
Notes
The script is part of the Avamar Administrator Windows installer. Default location is C:\Program Files\avs\administrator\bin. If the command fails, mcgui_login.bat also outputs to stdout the one (1) numeric return code follow by a specific event code and event text summary. For example:
1,22801,User login failure.
Return Codes
The mcgui_login.bat DOS batch file will always return one of the following numeric codes: 0 1 Command succeeded. Command failed.
307
Event Codes
21007 21008 21009 22801 23001 23996 GUI_INTERNAL GUI_VERSION_MISMATCH GUI_INVALID_PORT SEC_LOGIN_FAILURE CLI_MISSING_REQUIRE_ARGS CLI_CONNECT_FAILURE Internal Administrator client error. Version mismatch. Invalid port (version 3.1.0 only). User login failure. CLI command is missing required arguments. CLI failed to connect to MCS.
Examples
In all of the following examples, MCS-USER-ID must be your actual Avamar administrative user account, MCS-PSSWD is the login password for that account and avamar-1 is an example MCS name as defined in DNS. This example shows how to call mcgui_login.bat from within another script; the return value of zero (0) shows that the command successfully completed: cd C:\Program Files\avs\administrator call bin\mcgui_login.bat c:\jre1.5.0 MCS-USER-ID MCS-PSSWD / avamar-1 echo %ErrorLevel%
0
This example shows how to execute mcgui_login.bat from a Windows command prompt; the return value of one (1) and the error code information shows that the command did not successfully complete because of a user login failure: bin\mcgui_login.bat c:\jre1.5.0 MCS-USER-ID MCS-PSSWD / avamar-1
1,22801,User login failure.
308
mcserver.sh
The mcserver.sh shell script performs maintenance of the MCS.
Synopsis
mcserver.sh {--flush [NAME] | --help | --init | ---installcron | --ping | --publishevent --code=NUM --message=TEXT | --restart | --restore | --start | --status | --stop | --update} [--ap=PASSWORD] [--flushfile=FILE] [--flushtype=local] [--force] [--hfsaddr=AVAMARSERVER] [--hfsport=PORT] [--id=USER-ID] [--label=NAME] [--labelnum=NUM] [--local_hfsaddr=IP-ADDR] [--mcuserap=PASSWORD] [--newhfsaddr=IP-ADDR] [--newhfsport=PORT] [--oninstallcron] [--password=PASSWORD] [--restoretype={new-system | rename-system}] [--rootap=PASSWORD] [--server=AVAMARSERVER] [--test] [--xml]
Commands
One (and only one) of the following commands must be supplied on each command line: --flush [NAME] Backs up (flushes) MCS data. If NAME is supplied, that name will applied to the current backup. NOTE: The MCS automatically flushes 26 times per day (hourly and before each system checkpoint). --help --init Shows full online help (commands, options and full descriptions) for this utility, then exits. Initializes the MCS. IMPORTANT: Re-initializing a running MCS is highly destructive it will completely overwrite any custom preference settings stored in the live mcserver.xml file and revert the system state to the default settings. There is no way to recover your custom preference settings once they have been overwritten. --installcron Installs the dpn user crontab file. IMPORTANT: The MCS must already be running in order for command to successfully complete. This command added in version 3.7. --ping Performs a check (ping) to determine if the MCS is listening. This command added in version 3.0.
309
mcserver.sh COMMAND REFERENCE --publishevent --code=NUM --message=TEXT Force publishes an event with the specified numeric code and message text. IMPORTANT: The --code and --message options specify the numeric event code and event code message text, respectively; they must be supplied. This command and these options added in version 3.7. --restart Shuts down all administrator services except the PostgreSQL database (mcdb) and restarts them. Each restart is logged. Replaces all MCS data with data from the last flush (backup). If either --labelnum=NUM or --label=NAME is supplied, data from that backup will used for the restore. IMPORTANT: The MCS must be in a stopped state in order to do a restore. --start --status --stop --update Starts the MCS and logs the action. Returns status of each administrator service and any available performance statistics. Shuts down the administrator services and PostgreSQL database (mcdb). Each stop is logged. Upgrades database schema. mcserver.sh --update can only be run when the MCS is not running and will start up and shut down the database as needed to perform the upgrade.
--restore
Options
--ap=PASSWORD Used with --restore to specify MCUser password. Same as --password=PASSWORD. This option added in version 1.2.0. --flushfile=FILE Used with --restore to specify a local flush FILE. This option added in version 3.0. --flushtype=local Used with --flush and --restore commands to create or restore from a local MCS flush file, respectively. The mcserver.sh --flush --flushtype=local command generates a gzipped tar file with all the same files that normally get backed up during a normal (avtar) MCS flush. This tar file can be used to restore the MCS system state for troubleshooting and diagnostic purposes. This option added in version 3.0.
310
mcserver.sh COMMAND REFERENCE --force IMPORTANT: EMC internal use only. Used with --init, --restore and --update to bypass user prompts (force execution without regard for user inputs). This option added in version 1.2.0. --hfsaddr=AVAMARSERVER Used with --restore to specify which AVAMARSERVER IP address or fully qualified hostname (as defined in DNS) will be restored. Same as --server=AVAMARSERVER. This option added in version 1.2.0. --hfsport=PORT Used with --restore to restore from the Avamar server at this data PORT. This option added in version 3.0. --id=USER-ID Used with --restore to specify an alternate user account (USER-ID); that is another user account besides the default MCuser account. This option added in version 3.0. --label=NAME Used with --flush to specify a descriptive label identifying the contents of the flush. Used with --restore to specify which existing MCS backup (flush) should be used for the restore operation. --labelnum=NUM Used with --restore to specify which existing MCS backup (flush) should be used for the restore operation. Used with --restoretype=new-system to specify a new local Hash Filesystem (HFS) IP address (IPADDR) that will be used to store MCS preferences. This option added in version 3.0. --mcuserap=PASSWORD --newhfsaddr=IP-ADDR Used with --restore to specify a new Avamar root password. Used with --restoretype=new-system to specify a new HFS IP address (IP-ADDR) that will be used to store MCS preferences. This option added in version 3.0. --newhfsport=PORT Used with --restoretype=new-system to specify a new data PORT that will be used to store MCS preferences. This option added in version 3.0.
--local_hfsaddr=IP-ADDR
311
mcserver.sh COMMAND REFERENCE --noinstallcrom Used with --start to specify that dpn user crontab should not be installed. This is not the default setting. This option added in version 3.7. --password=PASSWORD Used to specify MCUser password when performing a restore. Same as --ap=PASSWORD. This option added in version 1.2.0. --restoretype= {new-system | rename-system} Used with --restore to specify the type of restore operation. If --restoretype is not supplied, a normal (nondisaster recovery) restore is performed. --restoretype=new-system is used to support system upgrades and disaster recovery in cases where a source Avamar server has been completely replicated to a destination server, and the destination server will be brought online to replace the old source server, or if you are using the source Avamar server as a swing box. --restoretype=rename-system when the Avamar server hostname or IP address has been changed. This option added in version 1.2.1. IMPORTANT: Beginning with version 2.1, support for dr and migrate restore types has been discontinued in favor of new-system and rename-system, respectively. --rootap=PASSWORD --server=AVAMARSERVER Used with --restore to specify a new MCUser password. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Same as --hfsaddr=AVAMARSERVER. This option added in version 1.2.0. --test --xml Returns MCS status (up or down). Used with --status to output an XML document instead of the standard --status output. This option added in version 2.0.0.
Notes
On Avamar 1.1.x and earlier systems, you must change directory to /usr/local/avamar/bin before running mcserver.sh. It will fail if it is run from any other location.
312
mcserver.sh COMMAND REFERENCE Flush Selection Logic. When an Avamar server has been populated with backups and MCS data (for example, group definitions, schedules, datasets, and so forth) from an existing system as part of a disaster recovery or upgrade operation (that is,the server has been the target of a root-to-root replication), it is especially important to restore the MCS state from the latest replicated flush; not from a local MCS flush. Failure to do so will cause replicated MCS data to be lost. Prior to version 3.0, when restoring the MCS following a server replication (by specifying --restoretype=new-system), it was necessary to manually examine a list of MCS flushes to determine which one (usually the latest remote flush) should be used for the restore procedure. Beginning with version 3.0, mcserver.sh now incorporates logic that automatically examines the list of flushes and selects that most recent one that is not from the local MCS. If no flushes from a replicated system are found, the following error message appears and the restore operation is terminated:
ERROR: No flushes from replicated Avamar server available. Make sure the source Avamar server has successfully replicated to MyServer.example.com.
Additionally, if a specific flush is specified using the --label or --labelnum options, mcserver.sh also verifies whether or not this is a replicated flush. If it is not, the following error message appears and the restore operation is terminated:
ERROR: The specified flush labelnum = 273 is not from the replicated Avamar server. You must restore a flush from the replicated Avamar server.
When performing a normal restore (that is, when not specifying --restoretype=new-system), the list of backups is similarly examined, and the most recent flush from the local MCS is used. If a specific flush is specified using the --label or --labelnum options then that particular flush is examined to determine if it will work for whatever type of flush was specified. When performing a normal restore, if there are no flushes from the local MCS, the following message appears:
WARNING: There are no flushes available for MyServer.example.com. Did you want to perform a "new-system" restore type Y/N?
If the specified flush is not from a local MCS, the following message appears:
WARNING: The flush you have selected to restore was replicated from a different Avamar server. Did you want to perform a "new-system" restore type Y/N?
In either case, if Y is chosen, a new-system restore is performed. If N is chosen, the restore operation is terminated.
Examples
To perform a local MCS flush using the default local flush filename and location (var/mc/server_flush/mcflush.YYYYMMDD.hhmmss), enter: mcserver.sh --flush --flushtype=local The following information appears in your command shell:
check.mcs passed === PASS === check.mcs PASSED OVERALL (poststart) Flushing MCS... MCS flushed to local file: var/mc/server_flush/mcflush.20050117.111101
To restore the MCS system state from the latest local flush file, enter: mcserver.sh --restore --flushtype=local
313
mcserver.sh COMMAND REFERENCE To restore the MCS system state from a specific local flush file (mcflush.20050117.111101 in this example), enter the following on a single command line: mcserver.sh --restore --flushtype=local --flushfile=mcflush.20050117111101
314
mcsmon_run
The mcsmon_run program runs monmcs (page 323) on the utility node. IMPORTANT: This program must be run as user dpn. This program attempts to load the dpnid OpenSSH key from the dpn user .ssh directory. This program will fail if it is run as a different user.
mcsnmp_cron
The mcsnmp_cron program is primarily intended to be run as a cron job in order to periodically restart the snmpd process and Avamar SNMP sub-agent. IMPORTANT: Documentation for this script is provided for reference purposes only. Do not run mcsnmp_cron directly from the command line.
315
mds_ctl
The mds_ctl program performs maintenance of the metadata search database. It is installed on the access node as part of the dpnmetadatadb rpm installation and requires no additional setup.
Synopsis
mds_ctl [--createlinks] [--dump] [--dumpfile=FILE] [--help] [--init] [--load] [--start] [--stop] [--update] [--version]
Options
--createlinks --dump --dumpfile=FILE Creates links to allow access to axionfs. Dumps the metadata search database to a file. Used with --dump to specify location of output file for the dump. Used with --load to specify which dump file to load. --help --init --load --start --stop --update --version Shows help, then exits. Initialize the metadata search database. Load the metadata search database from a file. Start postgres for the metadata search database. Stop postgres for the metadata search database. Update the metadata search database and preferences following an access node software upgrade. Shows version, then exits.
Notes
This program added in version 3.7. After installing any new version of the dpnmetadatadb rpm on an access node, run mds_ctl --update as user admin to perform any necessary updates.
316
metadata_cron
The metadata_cron program is primarily intended to be run as a cron job in order to perform indexing of the metadata search database (mdsdb) at regular intervals.
Notes
This program added in version 3.5.1.
317
mktime
The mktime program creates ntp.conf and step-tickers files for Avamar servers. IMPORTANT: Documentation for this script is provided for reference purposes only. The mktime program is typically invoked programmatically by asktime (page 43) and should not be run directly from the command line.
Synopsis
mktime [--configdir=DIR] [--debug] [--has-utility-node] [--help] [--mcs-zero-only] [--module-zero=id] [--single-node] [--verbose]
Options
--configdir=dir Specifies output directory (DIR) for time configuration files. Default location is /usr/local/avamar/var/time-config-files. Prints utility session information but does not actually perform the actions. Generates a configuration for a combined 0.s/0.m node. Shows help, then exits. Specifies that only the MCS in module zero be used. Specifies ID of module zero in a multi-node system. Generates single-node configuration. Provides maximum information.
Notes
The mktime program must be customized for each site, either by running asktime (page 43) or by manually editing a copy with a text editor such as vi or emacs. Instructions for manual modifications can be found at the top of the script. This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option. The mktime program also uses the SYSPROBEDIR environment variable (page 424) to determine where to place output files. This location can be overridden using the --configdir=DIR option.
318
Files
Files produced: ${configdir}/mktime.out ${configdir}/ntp.conf* ${configdir}/step-tickers* Where ${configdir} is one of the following, in order of decreasing precedence: 1. Specified by --configdir=DIR 2. ${SYSPROBEDIR} /time-config-files 3. /usr/local/avamar/var/time-config-files After running this script, run timedist (page 402) to install these files on the system, using SYSPROBEUSER=root or SYSPROBEUSER=dpn. Be sure to load the dpnid OpenSSH keys first. The asktime utility (page 43) does this automatically.
319
modify-snapups
The modify-snapups program writes a script to stdout, which when run, will either change backup expiration dates or delete any backups stored on the Avamar server with a creation date prior to the specified date.
Synopsis
modify-snapups --mode={delete | expire} [--after=DATE] [--before=DATE] [--days=NUM] [--domain=CLIENT-DOMAIN] [--help] [--include_mc_backups] [CLIENT-PATH ...]
Commands
One (and only one) of the following commands must be supplied on each command line: --mode={delete | expire} If --mode=delete is supplied, a script is written to stdout that will delete any backups stored on the Avamar server with a creation date prior to the date specified by --before=DATE. If --mode=expire is supplied, a script is written to stdout that will change backup expiration dates.
Options
--after=DATE If supplied with --mode=delete, all backups created after this date are eligible to be deleted. Default DATE setting in this mode is June 1, 1999 00:00:00. If supplied with --mode=expire, all backups created after this date are eligible to be expired. Default DATE setting in this mode is 90 days ago. --before=DATE If supplied with --mode=delete, all backups created before this date are eligible to be deleted. Default DATE setting in this mode is two weeks ago. If supplied with --mode=expire, all backups created before this date are eligible to be expired. Default DATE setting in this mode is now (the time at which modify-snapups is invoked). --days=NUM Used with the --mode=expire command to specify the new backup expiration date and time as the number (NUM) of whole days from today. Default value is 90 days from the initial backup creation date. If supplied, all clients within the specified CLIENTDOMAIN are processed.
--domain=CLIENT-DOMAIN
320
modify-snapups COMMAND REFERENCE --help Shows help, then exits. TIP: Specifying --mode=delete --help or --mode=expire --help provides context-sensitive help for those specific modes. --include_mc_backups If supplied, all clients within the special MC_BACKUPS domain are also processed. Because MC_BACKUPS is a special system domain, the default behavior is to not process these clients unless this option is explicitly supplied.
Client Path
CLIENT-PATH Specifies a valid Avamar client path (for example, /clients/MyClient). Additional client paths can be delimited with white space. IMPORTANT: Client paths must always terminate with a valid Avamar client name. You cannot specify a client domain in this manner. Use --domain when you want to process all clients within a particular domain. If not supplied, all clients are processed. This is the default setting.
Notes
IMPORTANT: This utility modifies backups but does not update the MCS database. This might cause the total bytes used value, which is used by the server licensing mechanism, to be incorrect. Therefore, EMC strongly suggests that you only use this utility if instructed to do so by EMC Technical Support. Contact EMC Technical Support for additional information. After modifying backups with this utility, you should immediately run dbUpdateactivitiesExt.pl (page 235) to update the MCS database. If you do not do this, your servers total bytes used value will be incorrect, which will adversely affect your server licensing (you will be deleting backups, which would normally free up additional storage capacity, but the licensing mechanism will be unaware that you have done so). The DATE specifier can be any date string acceptable to date(1). For GNU date(1), which on Linux is /bin/date, DATE can be just about any common date string, including such phrases as 2 weeks ago. Invoking modify-snapups --mode=delete is the same as running the delete-snapups utility (page 236). Invoking modify-snapups --mode=expire is the same as running the expire-snapups utility (page 275). This program added in version 3.0.
321
Examples
To create a default output script (one that will delete any backup stored /clients with a creation date older than two weeks from todays date) with the user-defined name del-old-backups, enter: modify-snapups --mode=delete /clients > del-old-backups To restrict program execution to specific client area (/clients/engineering in this example) and change the time period to three weeks ago, enter the following on a single command line: modify-snapups --mode=delete --before='3 weeks ago' /clients/engineering > del-old-backups After creating a script with the desired backup deletion parameters, enter the following to actually delete the backups from the Avamar server: /bin/sh -x ./del-old-backups
322
monmcs
The monmcs program obtains MCS status by running mcserver.sh --status.
Synopsis
monmcs [--mail]
Options
--mail If --mail is supplied, suspicious status issues cause mail to be sent to the admin user on the utility node.
morning_cron_run
The morning_cron_run program runs the morning cron job. morning_cron_run calls cp_cron (page 224), gc_cron (page 278) and hfscheck_cron (page 288) and is intended to be invoked by cron_env_wrapper (page 228). To run this script directly from a command shell, enter the following: cron_env_wrapper morning_cron_run IMPORTANT: This program must be run as user dpn. This program attempts to load the dpnid OpenSSH key from the dpn user .ssh directory. This program will fail if it is run as a different user.
323
nodenumbers
The nodenumbers program generates a table showing the logical node number, physical node number, IP address and MAC address of each node in an Avamar server. Refer to Physical vs. Logical Node Numbers (page 478) for additional information. The nodenumbers program compares the contents of the probe.out file (page 477) to the output of the avmaint nodelist command (page 122). This information is written to both STDOUT (screen) and to $SYSPROBEDIR/nodenumbers.out. The nodenumbers program requires that the Avamar server be running in order to deliver this information. This utility must be run as user admin.
Synopsis
nodenumbers [--hfsaddr=AVAMARSERVER] [--id=USER@AUTH] [--nodes=NODELIST] [--password=PASSWORD] [--probedir=PATH]
Options
--hfsaddr=AVAMARSERVER AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). This option added in version 1.2. --id=USER@AUTH Authenticate as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. This option added in version 1.2. --nodes=NODELIST Constrains output to a comma-separated LIST of nodes. This option added in version 1.2. --password=PASSWORD PASSWORD for the --id=USER@AUTH account. This option added in version 1.2. --probedir=PATH Specifies a PATH to a directory containing a valid probe.out file (page 477).
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
324
Output
Nodenumbers Utility [v1.6] Thu Oct 9 12:24:59 PDT 2003 Using /usr/local/avamar/var/probe.out Running 'avmaint nodelist --hfsaddr=10.0.30.10'. Appending to /usr/local/avamar/var/nodenumbers.out HFSCreateTime=1034287466 (Thu Oct 10 22:04:26 2002 UTC) Avamar Logical Node 0.0 0.1 0.2 0.3 0.4 0.5 0.6 1.0 1.1 1.2 1.7 < 1.8 1.9 1.A probe.out Physical Node 0.0 0.1 0.2 0.3 0.4 0.5 0.6 1.0 1.1 1.2 1.3 1.5 < 1.4 < 1.6 <
IP Address 10.0.30.10 10.0.30.11 10.0.30.12 10.0.30.13 10.0.30.14 10.0.30.15 10.0.30.16 10.0.31.10 10.0.31.11 10.0.31.12 10.0.31.13 10.0.31.17 < 10.0.31.15 < 10.0.31.18 <
MAC Address 00:30:48:51:CF:C8 00:30:48:51:D5:3F 00:30:48:51:D3:7E 00:30:48:51:EB:12 00:30:48:51:5D:5C 00:30:48:51:D5:F6 00:30:48:51:EF:3C 00:30:48:51:CF:CC 00:30:48:51:69:4E 00:30:48:51:D3:1E 00:30:48:51:D1:70 00:30:48:51:D1:82 00:30:48:51:B9:0F 00:30:48:51:E9:6E
Notes: - "<" means Out of Sequential Order. - "Physical" means "probe order", not rack location.
325
opstatus.dpn
The opstatus.dpn program returns Avamar server operational status as one of the following messages or exit codes:
EXIT CODE
MESSAGE
server up: ITEMS server degraded: ITEMS server down: ITEMS server unresponsive
0 1 2 3
Synopsis
opstatus.dpn [--help] [--quiet]
Options
--quiet --help If specified, only exit codes are returned. Shows help, then exits.
Notes
A status of server unresponsive means that opstatus.dpn could not contact the Avamar server for status. This might mean that there are network problems, or that the server non-operational on two or more nodes. When this status is returned, it is not possible to determine with certainty whether the server is operational or not. Further investigation of the storage nodes is required in order to make this determination.
326
pingmcs
The pingmcs program pings the MCS.
Synopsis
pingmcs [--mail]
Options
--mail If --mail is supplied, suspicious status issues cause mail to be sent to the admin user on the utility node.
327
probe
The probe program generates a probe.out file (page 477), which is used by other utilities to resolve simple MODULE.NODE designations into actual IP addresses.
Synopsis
probe [--admin=IP-ADDR] [--help] [--mcs=IP-ADDR] [--nat | --nonat] [--nodes=NUM] [--probedir=PATH] [--single] AVAMARSERVER Where AVAMARSERVER is the Avamar server hostname as defined in corporate DNS.
Options
--admin=IP-ADDR --help --mcs=IP-ADDR --nat | --nonat Specifies MCS IP address (IP-ADDR). Shows help, then exits. Specifies MCS (formerly known as the mcs) IP address (IP-ADDR). If --nat is supplied, probe knows that the Avamar server is configured to use Network Address Translation (NAT). If --nonat is supplied, probe knows that the Avamar server is not configured to use NAT. This is the default setting. --nodes=NUM --probedir=PATH --single Specifies number (NUM) of nodes to probe. Default value is all. Specifies directory PATH for output files. Creates a probe.out file for a single-node server. IMPORTANT: This option must be supplied when running probe on single-node systems.
Notes
By default, probe.out is saved in the /usr/local/avamar/var directory. If the SYSPROBEDIR environment variable (page 424) is set, then probe.out is saved to that location. When running probe on a multi-node server, the probe.out file will not be correctly generated if the utility node or any of the storage nodes are not functioning correctly. Additionally, probe requires a valid DHCP leases file in order for the storage nodes to correctly report their IP addresses. Once a probe.out file is created, it need not be changed unless nodes are added or removed from the Avamar server.
328
probeaux
The probeaux program is an auxiliary script used by probe (page 328). IMPORTANT: Documentation for this script is provided for reference purposes only. Do not run probeaux directly from the command line.
329
probedump
The probedump program reports storage node physical node IDs and their corresponding IP addresses, as found in the probe.out file. This program does not require that the Avamar server be running.
Synopsis
probedump [--help] [--nodes=NODELIST] [--probedir=PATH]
Options
--help --nodes=NODELIST --probedir=PATH Shows help, then exits. Constrains output to a comma-separated LIST of nodes. Specifies a PATH to a directory containing a valid probe.out file (page 477).
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
Example Output
[probedump v1.2] Using /usr/local/avamar/var/probe.out Physical Node IP Address 0.0 10.0.30.10 0.1 10.0.30.11 0.2 10.0.30.12 0.3 10.0.30.13 0.4 10.0.30.14 0.5 10.0.30.15 0.6 10.0.30.16 1.0 10.0.31.10 1.1 10.0.31.11 1.2 10.0.31.12 1.3 10.0.31.13 1.4 10.0.31.15 1.5 10.0.31.17 1.6 10.0.31.18 Note: - "Physical" means "probe order", not rack location.
330
probesingle
The probesingle program generates a probe.out file (page 477) on single-node servers only. It is functionally equivalent to running probe --single (page 328).
Synopsis
probesingle [--help] [--probedir=PATH] AVAMARSERVER Where AVAMARSERVER is the Avamar server IP address or hostname as defined in corporate DNS.
Options
--help --probedir=PATH Shows help, then exits. Specifies directory PATH for output files.
Notes
By default, probe.out is saved in the /usr/local/avamar/var directory. If the SYSPROBEDIR environment variable (page 424) is set, then probe.out is saved to that location.
331
propagate-gsan
The propagate-gsan program copies the correct version gsan binary to all storage nodes in the system in the same manner as restart.dpn --copy (page 364). However, propagate-gsan accomplishes this without restarting the data server and performs some additional checks. IMPORTANT: Documentation for This utility is provided for reference purposes only. The propagate-gsan program is typically invoked by dpnctl (page 239) and should not be run directly from the command line.
Synopsis
propagate-gsan [--debug] [--exedir=PATH] [--help] [--profiling] [--verbose]
Options
--debug --exedir=PATH --help --profiling --verbose Prints program session information but does not actually perform the actions. Specifies full PATH to parent directory of relevant executables. Shows help, then exits. If supplied, use profiling version of gsan binary. Provides maximum information.
332
psregrep
The psregrep program implements specialized regular expression (regex) pattern matching of Unix ps command output. It is primarily used to assist EMC engineering with debugging.
Synopsis
psregrep [--debug] [--help] [--pattern=REGEX] [--verbose]
Options
--debug --help --pattern=REGEX --verbose Prints program session information but does not actually perform the actions. Shows help, then exits. Same as supplying the help command. Search for this regular expression (REGEX) in ps output. Provides maximum information.
333
pull-checkpoint
The pull-checkpoint program retrieves (pulls) selected checkpoints tar bundles that were previously saved using store-checkpoint (page 397).
Synopsis
pull-checkpoint [--bump=N] [--debug] [--dryrun] --dst=NODE-LIST [--dstkey=FILE] [--gunzip] [--help] [--srckey=FILE] --src=NODE [--verbose]
Options
--bump=N --debug --dryrun --dst=NODE-LIST --dstkey=FILE --gunzip --help --src=NODE --srckey=FILE --verbose Increments (bumps up) source partition numbers by N. Prints utility session information but does not actually perform the actions. Runs auxiliary scripts in debug mode. Specifies destination Avamar nodes (NODE-LIST). This option is required. Specifies path to destination Avamar server OpenSSH dpnid key FILE. Un-gzips compressed tarballs (.tgz files). Shows help, then exits. Specifies source Avamar NODE. This option is required. Specifies path to source Avamar server OpenSSH dpnid key FILE. Provides maximum information.
Notes
IMPORTANT: You must load the dpnid OpenSSH key before running this utility. This program requires that source and destination Avamar server nodes have similar data partitioning schemes (for example, both source and destination servers have /data01 thru /data04 partitions).
334
Examples
Although some of these examples continue (wrap) to more than one line, all commands and options must be entered on a single command line (no line feeds or returns allowed). This example retrieves (pulls) available checkpoint bundle from source node 0.0 and unpacks it on destination nodes 0.2 to 0.6: pull-checkpoint --dst=0.2,0.3,0.4,0.5,0.6 --src=0.0 This example retrieves (pulls) available checkpoint tar bundle from the specified single-node server, using that server's SSH key: pull-checkpoint --dst=#.# --src=dpe17 --srckey=$HOME/.ssh/dpe17-dpnid
335
rebuild.node
The rebuild.node program rebuilds and restarts a node that has lost some or all of its content. First, the new node is restarted with the node ID of the non-operational node. Then, attempts are made to rebuild that node's content in place. If this is successful, this process is an effective alternative to attaching a node (with a new node ID), then decommissioning the old node.
Synopsis
rebuild.node --nodes=MODULE.NODE [--checkpointdir=DIR] [--checkpoints={all | CP-ID,CP-ID,...}] [--clean] [--copy] [--error] [--expert] [--force] [--hfsdir=PATH] [--kill] [--lmaddr=HOSTNAME] [--lmport=PORT] [--newdatacenter] [--norun] [--probedir=PATH] [--quiet] [--ramfsoptions=OPTIONS] [--runlevel={admin | fullaccess}] [--tag=MODULE.NODE] [--useram] [--usestunnel] [--verbose]
Node Descriptor
You must include the --nodes=MODULE.NODE node descriptor or the utility will not run, where MODULE.NODE is the physical node (page 478) you want to rebuild. Currently, only one node can be specified on each command line.
Options
--checkpointdir=DIR --checkpoints={all | CP-ID,CP-ID,...} Specifies the directory where the data is on each /data0? mount. Default value is cur. Used with --clean to completely remove one or more specified checkpoints (CP-IDs) or all checkpoints from the specified node. This option added in version 4.1. --clean If this option is not supplied (default behavior), rebuild.node checks storage nodes for any existing data and ensures that it is not overwritten (cleaned) during startup. However, if you want to rebuild with a clean storage node (all existing data removed), you must supply the --clean option. --copy --error --expert --force Copies files to the storage nodes. This is the default setting. Generates error when execution fails. Allows extra parameters to be passed through to storage nodes. This is not the default setting. Forces the server to be run even if there are stripes present on the node. This is not the default setting.
336
rebuild.node COMMAND REFERENCE --hfsdir=PATH Specifies the root location where the /data0? data mounts are located on the storage nodes. Default value is '/'. Terminates (kills) any running Avamar processes before starting. This is not the default setting. Specifies the login manager host address. Specifies the login manager PORT number. Specifies that this new node is in a new module (datacenter) that is being created. Does not actually execute commands. Specifies a PATH to a directory containing a valid probe.out file (page 477). Runs quietly. Controls stripes are placed in RAM, where OPTIONS is some combination of HWUG+XDCP designators, where: H W U G X C D P HFS stripes. Persistent store stripes. User accounting stripes. Delete stripes. Index stripes. Composite stripes. Data stripes. Parity stripes.
Multiple entries must be separated by commas. Default value is H+X,U+XP,G+P. --runlevel={admin | fullaccess} Sets server run level, to one of the following: admin Administration-only mode (only root and aroot users can access the server). Full access by all users.
fullaccess --systemname=NAME
Sets initial user-defined descriptive system name at server startup. NAME can be any combination of printable characters up to 32 characters in length.
337
rebuild.node COMMAND REFERENCE --tag=MODULE.NODE Specifies an optional logical node ID (MODULE.NODE) so that nodes can be rebuilt even if their logical ID is different from the physical ID. MODULE and NODE are both single-digit integers. For example, utility nodes are typically assigned 0.0 logical ID. --useram If supplied, RAM-resident stripes feature is enabled. This is not the default setting. This option added in version 4.0. --usestunnel Specifies that stunnel should be enabled when node is rebuilt. This is the default setting. Refer to Stunnel (page 35) for additional information. --verbose Provides maximum information.
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
--exedir=PATH
Dynamically locates the gsan executable according to the following rules: 1. If --exedir is supplied, use that gsan executable without further checking. 2. If gsan executable exists in the current working directory, use it. In this case, check if the gsan executable exists in /usr/local/avamar/bin also and, if so, print a warning if ./gsan is older than /usr/local/avamar/bin/gsan. 3. If gsan executable exists in /usr/local/avamar/bin, use it. 4. Otherwise, print a warning and set the exedir to current working directory (.).
338
rebuild.node COMMAND REFERENCE --freezelimits="STRING" Specifies upper limits for various processor activity statistics. The statistics must stay at or below the specified limits in order to consider the server to be in an idle state. STRING must be in the following format (which is also the default setting): busy=3,diskio=40,interrupt=200,switch=10 00 busy statistic refers to the percentage of time that the CPU is busy (the opposite of being idle); diskio refers to disk reads or writes per second; interrupt refers to interrupts per second; switch refers to context switches per second.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --useramfs IMPORTANT: Deprecated in version 4.0. Specifies whether or not to use ramfs feature. If specified, ramfs will be initialized and the server informed of its location. Default location is /ramroot. IMPORTANT: The --useramfs option is sticky. Once it has been specified, it will continue to be used after restarts or when starting new nodes. To turn off ramfs, you must use supply the --nouseramfs option. This option added in version 3.6.
Notes
IMPORTANT: Do not use both start.nodes (page 391) and rebuild.node to recover the same node. Doing so will cause the server to fail and might result in loss of data. This program added in version 3.6. This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
339
repl_cron
The repl_cron program is primarily intended to be run as a cron job during the evening_cron_run (page 274). The purpose for this script is to implement replication. The repl_cron program reads all input from the /usr/local/avamar/etc/repl_cron.cfg configuration file (page 481). A sample repl_cron.cfg is provided in the /usr/local/avamar/bin directory. In order to use replication, a live customized version must be placed in /usr/local/avamar/etc. This program added in version 1.2.
340
replicate
The replicate program recursively copies data from a source Avamar server to a destination Avamar server. Typically, replicate is invoked by way of repl_cron (page 340). repl_cron performs some housekeeping functions, then invokes replicate using settings stored in the repl_cron.cfg configuration file (page 481). repl_cron.cfg is a flag file that contains various commands, options and settings that are passed to the replicate utility at run time.
Synopsis
replicate --dstaddr=DEST-SERVER --dstid=USER@PATH --dstpassword=PASSWORD --hfsaddr=AVAMARSERVER --id=USER@AUTH --password=PASSWORD [{--fullcopy | --reportonly | --restore}] [--dpnname=SRC-SERVER | --dstpath=DOMAIN --srcpath=DOMAIN] [--after=TIMESTAMP] [--before=TIMESTAMP] [--dstport=PORT] [--exclude=PATTERN[,PATTERN...]] [--expiredelta=DAYS] [--flagfile=FILE] [--help] [--include=PATTERN[,PATTERN...]] [--pluginid-list=LIST] [--progresslog=FILE] [--retention-type={daily | weekly | monthly | yearly | none}] [--throttle=MBPS] [--timeout=SEC] [--version]
Required Values
In order to support automated nightly replication, the --dstaddr, --dstid and --dstpassword values are typically specified in the repl_cron.cfg configuration file (page 481) on the source server. The --hfsaddr, --id and --password values are typically read from the source server usersettings.cfg file (page 481). Therefore, these values do not typically have to be included in the repl_cron.cfg configuration file (page 481) or on the command line. --dstaddr=DEST-SERVER --dstid=USER@PATH Specifies a fully qualified DNS name or IP address of the destination Avamar server (DEST-SERVER). Authenticate on the desination Avamar server as this Avamar user ID (account name) at this domain. Where USER is the Avamar user name and PATH is the Avamar server domain. For example, supplying --dstid=admin@/REPLICATE/avamar-1, specifies that the admin user account in the destination server /REPLICATE/avamar-1 domain should be used to authenticate this replication session. This is useful for allowing selective replication on a domain-by-domain basis.
341
replicate COMMAND REFERENCE --hfsaddr=AVAMARSERVER --id=USER@AUTH Source AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Authenticate on the source Avamar server as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --password=PASSWORD PASSWORD for the --id=USER@AUTH account.
Commands
Commands control which operational mode (normal, full replicate, report-only or restore) is used. If no command is supplied, replicate operates in normal mode. If --fullcopy, --reportonly or --restore is supplied, replicate operates in full replicate, report-only or restore modes, respectively. One (and only one) of the following commands must be supplied on each command line: --fullcopy Asserts full root-to-root replication operational mode. Full replication is a special operational mode that creates a complete logical copy of an entire source server on the destination server. Furthermore, the replicated data is not copied to the REPLICATE domain, it is added directly to the root domain just as if source clients had registered with the destination server. Also, source server data replicated in this manner is fully modifiable on the destination server. --reportonly Asserts report-only operational mode. The mode is used to pre-determine the amount of storage a replication activity might consume on a destination server by running the replication job without actually saving any data to the destination server. This command added in version 4.1. --restore Asserts restore operational mode. If you previously replicated a source Avamar server to a destination Avamar server, running replicate from the destination server and supplying this command will restore that data to the source Avamar server.
342
Destination Descriptors
Destination descriptors are options that control where replicated data is stored on the destination server. --dpnname=SRC-SERVER Specifies a name that will be used to represent the source server (SRC-SERVER) on the destination server. This is the name that is used in the REPLICATE domain on the destination server. This option cannot be used with the --dstpath=DIR or --srcpath=DIR options. --dstpath=DOMAIN Specifies a location (DOMAIN) on the destination Avamar server where replicated source data will be stored. Default value is the top-level directory (/), which will store the replicated data in a new domain named for the source Avamar server. This option must be used in conjunction with the --srcpath=DOMAIN option and cannot be used with the --dpnname=SRC-SERVER option. --srcpath=DOMAIN Specifies a location (DOMAIN) on the source Avamar server from which to begin replication. Only data beneath this location is replicated. Default value is the top-level domain (/), which will replicate the entire server. This option must be used in conjunction with the --dstpath=DOMAIN option and cannot be used with the --dpnname=SRC-SERVER option.
Options
--after=TIMESTAMP Specifies that only backups matching TIMESTAMP and later should be replicated. TIMESTAMP must be specified using 24 hour local time zone values conforming to the following syntax: YYYY-MM-DD HH:MM:SS Partial date strings are permitted. For example, 200702 is equivalent to 2007-02-01 00:00:00. This option can also be used with --before=TIMESTAMP to define a range of effective dates. Only backups taken within this date range will be replicated. NOTE: Backups taken exactly at this TIMESTAMP will be replicated. This option added in version 3.5.
343
replicate COMMAND REFERENCE --before=TIMESTAMP Specifies that only backups taken before TIMESTAMP should be replicated. TIMESTAMP must be specified using 24 hour local time zone values conforming to the following syntax: YYYY-MM-DD HH:MM:SS Partial date strings are permitted. For example, 200702 is equivalent to 2007-02-01 00:00:00. This option can also be used with --after=TIMESTAMP to define a range of effective dates. Only backups taken within this date range will be replicated. This option added in version 3.5. --dstport=PORT Specifies data PORT to use when connecting to the destination Avamar server. Default is 27000. This option added in version 3.6. --exclude=PATTERN Excludes clients containing PATTERN from a replication. PATTERN is a single matching pattern. Common glob operators (wildcards) such as asterisk (*) and question mark (?) are allowed. Refer to Wildcards (page 39) for additional information. NOTE: When specifying exclusions, case sensitivity will vary according to the target computing platform you are backing up. Exclusions specified for Windows platforms are not case-sensitive; exclusions specified for most other platforms are case-sensitive. For example, --exclude=spot will exclude any source path name that contains the pattern spot. --exclude=/clients/ will exclude all clients within the /clients domain. Multiple patterns can be separated by commas (for example, --exclude=spot,/clients/). Multiple --exclude options can also be used to specify more than one PATTERN. --expiredelta=DAYS Causes the specified number of DAYS to be added to the expiration time for all newly replicated backups that already have expiration dates. Backups that do not have expiration dates are not affected (they still will have no expiration date after replication). Therefore --expiredelta=30 will add 30 days to the expiration dates of replicated backups that have expiration dates. --flagfile=FILE Specifies FILE containing a list of options and values that will be processed by this utility as if they were included on the command line. Default is /usr/local/avamar/etc/usersettings.cfg. Shows help for this utility, then exits.
--help
344
replicate COMMAND REFERENCE --include=PATTERN Includes clients containing PATTERN in a replication operation. PATTERN is a single matching pattern. Common glob operators (wildcards) such as asterisk (*) and question mark (?) are allowed. Refer to Wildcards (page 39) for additional information. NOTE: When specifying inclusions, case sensitivity will vary according to the target computing platform you are backing up. Inclusions specified for Windows platforms are not case-sensitive; inclusions specified for most other platforms are case-sensitive. For example, --include=spot will include any source path name that contains the pattern spot. --include=/clients/ will include all clients within the /clients domain. Multiple patterns can be separated by commas (for example, --include=spot,/clients/). Multiple --include options can also be used to specify more than one PATTERN.
345
replicate COMMAND REFERENCE --pluginid-list=LIST Constrains replication activity to only replicate backups originally taken with one or more plug-in types. LIST is a comma-separated list of one or more: Four-digit plug-in IDs Three-digit plug type designator Two-digit operating system designator This option added in version 4.1. Three-digit plug-in type designators: 000 001 002 003 004 005 006 007 008 009 011 012 avagent avtar filesystem Oracle/RMAN database NDMP Microsoft Exchange message Microsoft Exchange database Microsoft SQL Server database Script runner Replication DB2 database Microsoft Exchange 2007 database Microsoft Exchange 2007 web services
If a plug-in type designator is supplied, replication includes all backups taken with that plug-in type across all applicable operating systems. For example, supplying 009 includes all DB2 backups on both IBM AIX and Microsoft Windows operating systems.
346
replicate COMMAND REFERENCE --pluginid-list=LIST (continued) Two-digit operating system designators: 0 1 2 3 4 5 6 7 8 9 10 11 Unknown filesystem Linux Sun Solaris Microsoft Windows HP-UX IBM AIX MAC OSX Network Appliance Filer EMC Celerra Debugging Novell Netware FreeBSD
If an operating system designator is supplied, replication includes all backups originating on that specific operating system, regardless of which plug-in might have been used. For example, specifying 2 includes both Solaris filesystem backups as well as Oracle backups originating from databases hosted on Solaris platforms. --progresslog=FILE Specifies optional local session log FILE that can be used to monitor replication acivity progress. This option added in version 4.1.
347
replicate COMMAND REFERENCE --retention-type={daily | weekly | monthly | yearly | none} Constrains replication activity to only replicate backups assigned one of the following retention types: daily weekly monthly yearly none If supplied, only daily backups are replicated. If supplied, only weekly backups are replicated. If supplied, only monthly backups are replicated. If supplied, only yearly backups are replicated. If supplied, only backups without a specific replication type are replicated.
If not supplied, all backups, regardless of retention type, are replicated, subject to other filtering criteria such as --exclude, --include and so forth. Refer to Retention Types (page 31) for additional information about how backups are assigned retention types. This option added in version 4.0. --throttle=MBPS Controls rate at which the underlying avtar process sends data to the server. If --throttle=MBPS is supplied, avtar will pause as long as necessary after sending each packet in order to ensure that network usage does not exceed the specified maximum bandwidth; maximum bandwidth is specified in mega bits per second. For example, --throttle=5 uses half a 10Mbps connection, --throttle=0.772 restricts usage to one-half of a T1 link. --timeout=SEC Specifies a maximum execution time in seconds (SEC) for this replication session. The replicator process will be forcefully stopped if it has not successfully completed in this amount of time. Default time-out setting is 20 hours (72000 seconds). --version Returns the avtar and replicator software versions.
348
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them. If set false, only the most recent backup for each client is replicated. Default value is true. This option added in version 3.5. --avtar=OPTIONS --bindir=PATH --browse=DOMAIN --compress Specifies list of debugging options to be passed to avtar when it is invoked to perform the replication. Sets directory containing Avamar binary files. Default value is /usr/local/avamar/bin. Returns a list of sub-domains and clients within the specified server DOMAIN. Enables backup compression. This option added in version 4.0. --copysnapups If set false, new accounts are created on the destination server but no backups are copied. Default value is true. This option added in version 3.5. --count=MAX-NUM Specifies a maximum number (MAX-NUM) of backups to be replicated during each replication session. This option added in version 3.5. --diffaccts=LEVEL Specifying --diffaccts=0 disables the optimization, which compares the accounting database and applies only the differences. Controls whether replicate should recursively process. Specifying --dontfork disables the --timeout option. Specifies list of debugging options to be passed to avmaint when it is invoked to perform operations on the remote (destination) server. This option added in version 3.5. --dstavmgr=OPTIONS Specifies list of debugging options to be passed to avmgr when it is invoked to perform operations on the remote (destination) server.
--allsnapups
--dontfork
--dstavmaint=OPTIONS
349
replicate COMMAND REFERENCE --forcecreate Forces all accounts to be created on the destination server even when other options (for example, (for example, --include, --exclude, and so forth) are supplied. This option added in version 3.5. --globalcid If supplied, global client IDs (CIDs) are used during replications. Global CIDs are primarily used to facilitate fast failovers from one server to another following a full root-to-root replication. This is the default setting. This option added in version 3.7. --maxchunksize=BYTES Sets maximum size of an atomic chunk. Default value is 61440. This option added in version 4.0. --minchunksize=BYTES Sets minimum size of an atomic chunk. Default value is 1000. This option added in version 4.0. --nochunkconversion If supplied, automatic chunk conversion is not performed. This option added in version 4.0. --repldebug --rechunk=(disable | enable | default) Enables debugging mode in replicator. Controls whether or not replicated data should be rechunked in order to maximize data deduplication on the destination server. One of the following: disable enable Do not re-chunk data before storing on the destination server. Re-chunk data before storing on the destination server in order to maximize data deduplication. Automatically re-chunk data when source and destination server chunking parameters are different.
default
This option added in version 4.0. --small-client-mb=MB Threshold under which a client's new data is considered small. Default setting is 128MB of new data. A setting of 0 disables this optimization. This option added in version 3.7.
350
replicate COMMAND REFERENCE --srcavmaint=OPTIONS Specifies list of debugging options to be passed to avmaint when it is invoked to perform operations on the local (source) server. This option added in version 3.5. --srcavmgr=OPTIONS Specifies list of debugging options to be passed to avmgr when it is invoked to perform operations on the local (source) server. Sets atomic block sticky byte magic number (NUM). Default value is 30000. This option added in version 4.0. --usesnapuplist If supplied, logic that optimizes normal replications by examining backup lists is enabled. This is the default setting. This option added in version 3.7. --workdir=PATH Specifies alternate working directory for replicator temporary files. Default workdir is /tmp/replicate. This directory will be created if it does not exist.
--threshold=NUM
351
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged: --failover --centera IMPORTANT: Deprecated in version 3.7 in favor of --globalcid. IMPORTANT: Deprecated in version 3.7. Designates that this replication operation is from an EMC Centera external data repository. This option added in version 3.5.1.
Notes
To completely replicate the entire contents of a source Avamar server to the destination server, include --srcpath=/, --dstpath=/ and --fullcopy options. Also ensure that both --srcpath=/ and --dstpath=/ are set to root (/).
Examples
One-Time Client Replications. Often, when a large client is backed up to an Avamar server, the next replication session might run over the allotted time. Rather than permanently modifying replication timing, it is often easier to just explicitly include that large client on a replication command line and run it at your convenience. This example replicate shows how explicitly include a client (spot) as part of a one-time replication: nohup replicate --flagfile=/usr/local/avamar/etc/repl_cron.cfg --include=spot --timeout=0 --throttle=4 >& spot.replout & NOTE: Space limitations in this publication cause the previous replicate command to continue (wrap) to more than one line. Your command must be entered on a single command line (no line feeds or returns allowed). The order of the options is important because the --include and --timeout options override the --exclude and --timeout flags in repl_cron.cfg. The logging information will be written to spot.replout in the current working directory. This command should be run in the background with nohup so that you can log out of the system without interrupting this replication job.
352
resite
The resite program is an interactive utility used to change a single-node Avamar servers network configuration (for example, hostname, IP address, parent domain, and so forth). IMPORTANT: After running the resite program, you must manually update preferences files for other applications and features to reflect these changes. Applications and features known to be impacted by resite include EMS and Avamar Administrator Command Line Interface (CLI).
Synopsis
resite [--config=PATH] [--debug] [--help] [--ignore_filesystem_requirements] [--ignore_node_type] [--log=PATH} [--verbose]
Options
--config=PATH --debug --help --ignore_filesystem_requirements --ignore_node_type --log=PATH --verbose Specifies PATH to configuration file. Prints utility session information but does not actually perform the actions. Shows help, then exits. Does not check filesystem requirements. Does not check the node type. Specifies log file PATH. Provides maximum information.
Notes
The resite program must be run as the operating system root user. The resite program poses interactive questions as dialog boxes. Because cursor movement capabilities are required for the interactive questions, you should not run resite from an Emacs shell buffer. It is recommended that resite be invoked directly from the console of the Avamar server, using a directly attached keyboard and monitor. Although it is possible to invoke resite from an interactive ssh or PuTTY session, be aware that the session will drop when the network configuration changes. In that event, re-connect to the Avamar server at its new name or IP address.
353
resite COMMAND REFERENCE Interactive questions are bypassed if --config=PATH is specified and if the specified configuration file contains all of the required information. PATH should be specified as an absolute path (starting with a /). The following are example configuration file entries:
hostname=avamar-1 dns_domain_name=example.com eth0_ipv4_address=10.0.254.233 eth0_network_mask=255.255.255.0 default_gateway_address=10.0.254.1 primary_dns_resolver_address=192.168.100.89 secondary_dns_resolver_address=192.168.200.89 smtp_server=mail storage_server_name=AVAMARSERVER local_time_zone=US/Pacific
The resite program creates a default configuration file, /usr/local/avamar/var/resite.conf, if it does not exist. The following log file should be reviewed after running resite: /usr/local/avamar/var/resite.log Expect and ignore the following error messages in the log file:
database "mcdb" already exists (from MCS) User or account already exists (from EMS)
Other error messages, if they appear, should be scrutinized and addressed. A progress summary of a resite session can be found in /usr/local/avamar/var/resiteaux.out. resiteaux is the name of a background process that applies the network configuration changes.
Examples
The following example shows how to properly use the resite utility to change a single-node Avamar servers network settings: 1. Gather all the following new site settings: (a) New name of the host (for example, avamar-1) (b) New parent domain name for the host (for example, example.com) (c) New IP address for network interface eth0 (for example, 10.0.5.5) (d) New network mask for network interface eth0 (for example, 255.255.255.0) (e) New default network gateway address (for example, 10.0.5.1) (f) IP address of the primary DNS resolving name server (for example, 192.168.100.89) (g) IP address of the secondary DNS resolving name server (for example, 192.168.200.89) (h) Name, as defined in corporate DNS, of an email server that will accept outgoing (SMTP) emails from the Avamar server (for example, mail) (i) Name of the local time zone (for example, US/Pacific) (j) Hostname of the Avamar storage server (for example, AVAMARSERVER) AVAMAR 4.1 TECHNICAL ADDENDUM 354
2. Open a command shell. 3. Log into the server as user admin. 4. Load the admin OpenSSH key by entering: ssh-agent bash ssh-add ~admin/.ssh/admin_key You are prompted to enter a passphrase. 5. Enter the admin user account passphrase and press ENTER. IMPORTANT: You must force a flush of the MCS database in order to ensure that any very recent changes to schedules and other administrative settings are saved. Failure to do so will result in loss of data. 6. Enter: mcserver.sh --flush 7. Switch user to root by entering: su When prompted for a password, enter the root password and press ENTER. 8. Enter: resite The following information appears in your command shell:
Answer each of the remaining questions with information you collected in step 1.
355
resite COMMAND REFERENCE 9. Navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
10. Enter the new name of the host (for example, avamar-1), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
11. Enter the new parent domain name for the host (for example, example.com), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
356
resite COMMAND REFERENCE 12. Enter the new IP address for network interface eth0 (for example, 10.0.5.5), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
13. Enter the new network mask for network interface eth0 (for example, 255.255.255.0), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
14. Enter the new default network gateway address (for example, 10.0.5.1), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
357
resite COMMAND REFERENCE 15. Enter the new IP address of the primary DNS resolving name server (for example, 192.168.100.89), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
16. Enter the new IP address of the secondary DNS resolving name server (for example, 192.168.200.89), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
17. Enter the name, as defined in corporate DNS, of an email server that will accept outgoing (SMTP) emails from the Avamar server (for example, mail), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
358
resite COMMAND REFERENCE 18. Enter the new Avamar storage server name (for example, AVAMARSERVER), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
359
You are using Network Address Translation (NAT) at your site. You are not using Network Address Translation (NAT) at your site.
Enter the external IP address clients can use to contact the MCS, then navigate to the OK field using your cursor keys and press ENTER. Leave this field blank, then navigate to the OK field using your cursor keys and press ENTER.
20. Navigate to the proper time zone using your cursor keys and press ENTER; then navigate to the OK field using your cursor keys and press ENTER.
360
resite COMMAND REFERENCE 21. Enter the name, as defined in corporate DNS, of an email server that will accept outgoing (SMTP) emails from the Avamar server (for example, mail), then navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
22. Review your settings, then navigate to the OK field using your cursor keys and press ENTER. The following information might appear in your command shell:
361
resite COMMAND REFERENCE 23. Navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
24. Navigate to the OK field using your cursor keys and press ENTER. The following information appears in your command shell:
25. Navigate to the OK field using your cursor keys and press ENTER. After all questions have been answered, resite will begin making changes in the background. resite typically takes about 10-20 minutes to complete these changes. However, on very large systems, resite can take up to two hours to complete all required changes.
362
resite COMMAND REFERENCE If you invoked resite from an interactive ssh or PuTTY session, your session will drop when the network configuration changes take affect. If this occurs, it is usually possible to log back into the server using the new hostname or IP address even if all resite tasks have not completed. Doing so would allow you monitor /usr/local/avamar/var/resite.log in order to determine when resite has completed. 26. Enter: tail /usr/local/avamar/var/resite.log 27. Ensure that resite has completed without errors before resuming normal server operation.
363
restart.dpn
The restart.dpn program restarts Avamar processes on specified storage nodes in the Avamar server after the system has been shut down.
Synopsis
restart.dpn [--checkpointdir=DIR] [--copy] [--delay] [--error] [--expert] [--hfscheck] [--hfsdir=PATH] [--kill] [--mainhost=IP-ADDR] [--nodes=NODE-LIST] [--offlineok] [--probedir=PATH] [--quiet] [--ramfsoptions=OPTIONS] [--runlevel={admin | fullaccess}] [--savesysinfo] [--skipvalidation] [--sslport=PORT] [--useram] [--usestunnel] [--usewritelogheaders] [--verbose] [--wait]
Options
--checkpointdir=DIR --copy --delay Specifies the directory where the data is on each /data0? mount. Default value is 'cur'. Copies files to the storage nodes. This is not the default setting. Supplying --delay introduces a short pause between node restarts on multi-node servers. --nodelay is the default setting. This option added in version 3.0. --error --expert --hfscheck Generates error when execution fails. Allows extra parameters to be passed through to storage nodes. This is not the default setting. Restarts system to perform an HFS check. Refer to Checkpoint Validation (HFS Checks) (page 15) for additional information. Specifies the root location where the /data0? data mounts are located on the storage nodes. Default value is '/'. Terminates (kills) any running Avamar processes before starting. This is not the default setting. Sets mainhost storage node to this IP address (IPADDR). Default value is IP address of storage node 0.0. Specifies which node(s) to restart. If not supplied, all storage nodes are restarted. Requires physical MODULE.NODE designations as defined in probe.out (page 477). Refer to Physical vs. Logical Node Numbers (page 478) for additional information.
--hfsdir=PATH
--kill --mainhost=IP-ADDR
--nodes=NODE-LIST
364
restart.dpn COMMAND REFERENCE --offlineok --probedir=PATH --quiet --ramfsoptions=OPTIONS Allows off-line stripes. This parameter is passed to wait.dpn. Specifies a PATH to a directory containing a valid probe.out file (page 477). Runs quietly. Controls stripes are placed in RAM, where OPTIONS is some combination of HWUG+XDCP designators, where: H W U G X C D P HFS stripes. Persistent store stripes. User accounting stripes. Delete stripes. Index stripes. Composite stripes. Data stripes. Parity stripes.
Multiple entries must be separated by commas. Default value is H+X,U+XP,G+P. This option added in version 3.6. --runlevel= {admin | fullaccess} Sets server run level, to one of the following: admin Administration-only mode (only root and aroot users can access the server). Full access by all users.
fullaccess
IMPORTANT: Beginning with version 2.0.2, support for numeric run levels 4 and 6 has been discontinued in favor of using admin and fullaccess string names, respectively. --savesysinfo If supplied, special storage server (GSAN) cleanliness tests are run. This is the default setting. When supplied, server integrity checks are bypassed. This is not the default setting. Specifies SSL data PORT. Default is 29000. Refer to Stunnel (page 35) for additional information. This option added in version 3.6.
--skipvalidation --sslport=PORT
365
restart.dpn COMMAND REFERENCE --startmainhost IMPORTANT: Beginning with version 2.0.2, support for this option has been discontinued (you should not use it). Forces selection of mainhost storage node, even if only one or a few nodes are being restarted. --useram If supplied, RAM-resident stripes feature is enabled. This is not the default setting. This option added in version 4.0. --usestunnel Specifies that stunnel should be enabled when server is restarted. This is the default setting. Refer to Stunnel (page 35) for additional information. This option added in version 3.6. --usewritelogheaders Enables feature that periodically updates server writelog header in order to prevent writelog corruption. This is the default setting. This option added in version 3.7. --verbose --wait Provides maximum information. Calls wait.dpn (page 419) after starting the storage nodes. This is the default setting.
366
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
--exedir=PATH
Dynamically locates the gsan executable according to the following rules: 1. If --exedir is supplied, use that gsan executable without further checking. 2. If gsan executable exists in the current working directory, use it. In this case, check if the gsan executable exists in /usr/local/avamar/bin also and, if so, print a warning if ./gsan is older than /usr/local/avamar/bin/gsan. 3. If gsan executable exists in /usr/local/avamar/bin, use it. 4. Otherwise, print a warning and set the exedir to current working directory (.). This option added in version 3.0.
--freezelimits="STRING"
Specifies upper limits for various processor activity statistics. The statistics must stay at or below the specified limits in order to consider the server to be in an idle state. STRING must be in the following format (which is also the default setting): busy=3,diskio=40,interrupt=200,switch=1000 busy statistic refers to the percentage of time that the CPU is busy (the opposite of being idle); diskio refers to disk reads or writes per second; interrupt refers to interrupts per second; switch refers to context switches per second. This option added in version 3.5.
367
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --useramfs IMPORTANT: Deprecated in version 4.0. Specifies whether or not to use ramfs feature. If specified, ramfs will be initialized and the server informed of its location. Default location is /ramroot. IMPORTANT: The --useramfs option is sticky. Once it has been specified, it will continue to be used after restarts or when starting new nodes. To turn off ramfs, you must use supply the --nouseramfs option. This option added in version 3.6.
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option. It is usually sufficient to restart the Avamar server simply by running restart.dpn without any parameters as long as the following conditions are satisfied: 1. A valid probe.out file (page 477) is located in /usr/local/avamar/var or in the directory specified by the SYSPROBEDIR environment variable (page 424). 2. You have run ssh-add admin_key or are prepared to directly enter the admin password as the ssh command is programmatically run on each storage node. 3. The dpnutils RPM is installed. 4. The /usr/local/avamar/bin directory is in your path.
368
resume_crons
The resume_crons program resumes regular execution of any suspended Avamar server cron scripts, such as cp_cron (page 224), gc_cron (page 278) and hfscheck_cron (page 288). It does not affect any currently running scripts.
369
rollback.dpn
The rollback.dpn program returns (rolls back) an Avamar server to the last saved state in a checkpoint. IMPORTANT: This should only be used to roll back to a checkpoint that has passed an HFS check.
Synopsis
rollback.dpn --cptag=CP-ID [--copy] [--parallel] [--restart] [--savesysinfo] [--yes] In order to run rollback.dpn, you must know a valid checkpoint CP-ID to roll back to. To determine what the last good checkpoint is, you must look in the /usr/local/avamar/var/cron/hfscheck.log file. Go to the end of the file and search backwards for the string completed hfscheck of cp. For example:
======== completed hfscheck of cp.20021020160014 ========
Where cp.20021020160014 is the checkpoint tag you must supply with rollback.dpn.
Options
--copy --parallel --restart --savesysinfo --yes If --copy is supplied, files are copied to the storage nodes; this is the default behavior. If --parallel is supplied, server rollback of all nodes in done in parallel; this is the default behavior. If --restart is supplied, server is automatically restarted following the rollback; this is the default behavior. If supplied, special storage server (GSAN) cleanliness tests are run. This is the default setting. Does not prompt for confirmation to restart server if only one error occurs.
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
370
rununtil
The rununtil program is used by other Avamar programs to run a process for a specified amount of time. IMPORTANT: Documentation for this script is provided for reference purposes only. Do not run rununtil directly from the command line.
371
sched.sh
The sched.sh script returns a histogram representation of daily server activities. Each day is separated by a line break and multiple lines are shown for a single day when activities overlap. The heading columns show the time of day starting from midnight (12am). By default, each column represents 30 minutes.
Synopsis
sched.sh [--csv | --text] [--days=NUM] [--help][--wide]
Options
--csv | --text If --csv is supplied, output is formatted as Comma Separated Values (CSV), which can be easily copied and pasted into a spreadsheet. If --text is supplied, output omits special formatting characters, which makes it easier to copy, paste and reuse the information. If neither option is supplied, output is in the default format, which uses special formatting characters for enhanced clarity. This is the default. --days=NUM --help --wide Limits scope of report to only include the specified number (NUM) of days. Default is 7 days. Shows help, then exits. Changes time scale to 15 minute increments, resulting in output that is twice as wide. This is not the default.
Notes
Because of the inherent limitations of using text characters to represent a histogram, some routines may appear to overlap that do not. Individual log files should be examined to confirm exact times of all operations. This program added in version 4.1.
372
Examples
sched.sh
373
scn
The scn program is the Avamar secure file copy utility. It wraps the OpenSSH scp utility in order to accept simpler MODULE.NODE designations.
Synopsis
scn [--displaymap] [{--logical | --physical}] [--probedir=PATH] [--skipserver]
Options
--displaymap Displays a list of node designations and IP addresses currently assigned to them and exits. If --logical is specified, logical node designations are used as primary key. If --physical is specified, physical node designations in probe.out (page 477) are used as primary key. If this utility requests this information from the Avamar server and does not receive a list of current nodes, /usr/local/avamar/var/nodelist.xml is read as a backup source of this information. This option added in version 2.0. --logical | --physical If --logical is supplied, all subsequent node designations are assumed to be logical. If --physical is supplied, all subsequent node designations are assumed to be physical. --probedir=PATH --skipserver Specifies a PATH to a directory containing a valid probe.out file (page 477). If supplied, this utility does not request a current list of nodes, instead it use parses /usr/local/avamar/var/nodelist.xml for a current list of nodes. This option added in version 2.0.
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option. This utility reads a default operating system user ID from the SYSPROBEUSER environment variable (page 425). If SYSPROBEUSER is not set, then remote commands are run as user admin.
374
Examples
Consider a typical scp command line: scp admin@10.0.22.10:/data01/cur/gsan.opt . The equivalent scn command allows you to use MODULE.NODE syntax as follows: scn 0.0:/data01/cur/gsan.opt .
375
showperfhistory
The showperfhistory program runs the avmaint perf status (page 124) command and displays the average disk read performance rates in an easy-toview format, sorted first by event sets, then by average read rate.
Synopsis
showperfhistory [--noreset] [--zeroaverages]
Options
--noreset --zeroaverages Suppresses display of suggested avmaint perf reset (page 124) command and the leading comment character. Displays statistics even if the average is zero because the count is zero.
Notes
This program added in version 4.1.
Examples
showperfhistory
# # # # # # # # # # # # 0.0 0.2 0.0 2 0 0 772 775 775 52 38 49 2 1 2 2 2 2 118 197 118 4.46 37.00 42.00 17 backup,hfscheck 17 backup,hfscheck 17 backup,hfscheck # avmaint perf reset 0.0 --disknum=2 --events=17 0.1 0.0 0.2 0 0 1 785 775 771 39 49 42 1 2 1 10 10 10 2 2 2 72.88 74.03 75.15 1 backup 1 backup 1 backup 0.1 0.0 0.0 0 0 1 | 785 775 783 Per Disk 39 49 41 1 2 3 | 10 10 10 243 243 243 Per Disk and Event Set 72.73 72.84 73.23 0 0 0 # Node Disk | Used Skipped Failed | Days Count Average Event-Bits Events
Note that this example shows an avmaint perf reset command because showperfhistory detected an abnormally high average read performance value on node 0.0 disk 2. IMPORTANT: The showperfhistory program only displays a suggested avmaint perf reset command when it detects an unusually high or low value compared to other nodes or disks for the same set of events. It must be understood that this is only a suggested avmaint perf reset command based on simple heuristics. It is the administrators responsibility to determine if this value should actually be reset, then issue the command to do so.
376
showperfhistory COMMAND REFERENCE The following table describes each showperfhistory output column:
COLUMN DESCRIPTION
Logical node number. A particular disk on that node. Shows total number of performance monitoring tests run on that disk. Shows total number of performance monitoring tests whose results were discarded because the event set at the start of the test did not match the set at the end of the test. Shows total number of performance monitoring tests whose results were out of tolerance. Shows total number of days worth of statistics that have accumulated for a particular disk and set of events. Shows total number of tests that have been run over the specified number of days. Shows the average read performance in MB/sec. Shows numeric encoding for the set of events listed in the Events column. Describes which events took place.
Skipped
Failed
PER DISK AND EVENT SET
377
shutdown.dpn
The shutdown.dpn program shuts down the Avamar server. It wraps various avmaint (page 53) commands.
Synopsis
shutdown.dpn [--kill] [--nocheckhfs] [--nocheckserver] [{--now | --nowait}] [--pollinterval=SEC] [--usestunnel]
Options
--kill Explicitly terminates (kills) all running Avamar server (gsan) processes on all nodes by issuing the correct Linux killall commands at the correct time during the server shut down process. IMPORTANT: Unless --nokill is supplied, a shutdown.dpn --now command also terminates (kills) all running Avamar server (gsan) processes on all nodes. Therefore, shutdown.dpn --kill should only be used if the server does not shut down properly after issuing a shutdown.dpn --now command. --nocheckhfs --nocheckserver Turns off the check for currently active HFS check. Causes shutdown.dpn to not wait for all Avamar server processes to gracefully exit before performing the shutdown. If --now is supplied, an avmaint shutdown --now command is issued, which immediately cancels any active client sessions. If --nowait is supplied, active client sessions are shut down in the following manner: 1. Run avmaint suspend to lock out new avtar connections to the Avamar server. 2. Run avmaint sessions to check for active backups. 3. One of the following: If active backups found, re-enable dispatcher connections with avmaint resume and terminate script session. If no backups found, run avmaint shutdown. --pollinterval=SEC --usestunnel Sets the interval in seconds that avmaint sessions is successively called. Default value is 6 seconds. Specifies that stunnel should be shut down on all nodes when server is shut down. This is the default setting. Refer to Stunnel (page 35) for additional information. This option added in version 3.6.
--now | --nowait
378
Notes
Invoking shutdown.dpn with no options performs an avmaint suspend command to lock out new avtar connections to the Avamar server, then goes into a loop calling avmaint sessions until there are no more client sessions in progress. At any point during this loop, you can press CTRL+C to cancel the shutdown operation. This will also re-enable dispatcher connections with avmaint resume before exiting. When the number of client sessions goes to 0, the normal avmaint shutdown is run (which should be quick because there are no more clients to wait for). After the avmaint shutdown command is run, shutdown.dpn waits for all Avamar server (gsan) processes to end before it exits. If an HFS check is running, shutdown.dpn will not shut down the server.
379
ssn
The ssn program is the Avamar secure remote shell utility. This utility wraps the OpenSSH ssh utility in order to accept simpler MODULE.NODE designations.
Synopsis
ssn [--displaymap] [{--logical | --physical}] [--probedir=PATH] [--skipserver]
Options
--displaymap Displays a list of node designations and IP addresses currently assigned to them and exits. If --logical is specified, logical node designations are used as primary key. If --physical is specified, physical node designations in probe.out (page 477) are used as primary key. If this utility requests this information from the Avamar server and does not receive a list of current nodes, /usr/local/avamar/var/nodelist.xml is read as a backup source of this information. This option added in version 2.0. --logical | --physical If --logical is supplied, all subsequent node designations are assumed to be logical. If --physical is supplied, all subsequent node designations are assumed to be physical. --probedir=PATH --skipserver Specifies a PATH to a directory containing a valid probe.out file (page 477). If supplied, this utility does not request a current list of nodes, instead it use parses /usr/local/avamar/var/nodelist.xml for a current list of nodes. This option added in version 2.0.
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option. This utility reads a default operating system user ID from the SYSPROBEUSER environment variable (page 425). If SYSPROBEUSER is not set, then remote commands are run as user admin.
380
Examples
Consider a typical ssh command line: ssh admin@10.0.22.10 COMMAND Where COMMAND is the command set you want to execute on the remote node. The equivalent ssn command allows you to use MODULE.NODE syntax as follows: ssn 0.0 COMMAND
381
start.dpn
The start.dpn program starts an Avamar server for the first time and initializes all the storage nodes in the system. IMPORTANT: After start.dpn has been run once and the Avamar server has been initialized, start.dpn should not be run again. To re-start a server that has existing data, use restart.dpn (page 364).
Synopsis
start.dpn [--catserver] [--check] [--checkpointdir=DIR] [--clean] [--copy] [--delay] [--diskreadonly=PERCENT] [--diskwarning=PERCENT] [--encrypt={proprietary | ssl | ssl:AES128-SHA | ssl:AES256-SHA}] [--encryptatrest=ENCRYPTIONKEYSALT] [--error] [--exedir=PATH] [--expert] [--hfscheck] [--hfsdir=PATH] [--indexelements=NUM] [--kill] [--lmaddr=HOSTNAME] [--lmport=PORT] [--mainhost=IP-ADDR] [--masterdc=NUM] [--matchbits=N] [--maxlogfiles=NUM] [--maxlogsize=MB] [--maxrwatomdatastripe=SIZE] [--minstripestowrite=NUM] [--nat] [--noinit] [--norun | --n] [--nousercheck] --password=PASSWORD [--probedir=PATH] [{--quiet | -q}] [--paritygroups=Nx,Fy] [--ramfsoptions=OPTIONS] [--runlevel={admin | fullaccess}] [--rwmatchbits=N] [--savesysinfo] [--short] [--sslport=PORT] [--systemname=NAME] [--tag] --user=USER@AUTH [--useram] [--usermatchbits=N] [--usestunnel] [--usewritelogheaders] [{--verbose | -v}]
Options
--catserver Starts this server as a CAT server in order to perform data de-duplication assessment for prospective customer. Supplying --check runs check.dpn. This is the default setting. Specifies the directory where the data is on each /data0? mount. Default value is 'cur'. Calls hfsclean before starting. Default value is FALSE. Copies files to the storage nodes. This is the default setting. Supplying --delay introduces a short pause between node starts on multi-node servers. --nodelay is the default setting. This option added in version 3.0.
382
start.dpn COMMAND REFERENCE --diskreadonly=PERCENT Sets percentage (PERCENT) of full server storage capacity that will trigger conversion of the server from a fully writable condition to read-only. Default value is 65%. Sets percentage (PERCENT) of full server storage capacity that will trigger a warning to the server administrator that the server is becoming full. Sets encryption method. Valid settings are: proprietary ssl:AES128-SHA No encryption. 128-bit Advanced Encryption Standard. 256-bit Advanced Encryption Standard. A special mode in which start.dpn negotiates and uses the strongest encryption setting that the client can support.
--diskwarning=PERCENT
ssl:AES256-SHA
ssl
Default setting is proprietary. --encryptatrest=ENCRYPTIONKEYSALT Enables permanent encryption of all user data stored on the server. This feature uses blowfish encryption. ENCRYPTIONKEYSALT is a password or phrase used to generate the encryption key. IMPORTANT: Enabling this feature is a permanent decision that can only be made when a new server is started for the first time. Enabling this feature will affect overall server CPU usage and performance. This option added in version 3.5. --error Generates error when execution fails.
383
start.dpn COMMAND REFERENCE --exedir=PATH Dynamically locates the gsan executable according to the following rules: 1. If --exedir is supplied, use that gsan executable without further checking. 2. If gsan executable exists in the current working directory, use it. In this case, check if the gsan executable exists in /usr/local/avamar/bin also and, if so, print a warning if ./gsan is older than /usr/local/avamar/bin/gsan. 3. If gsan executable exists in /usr/local/avamar/bin, use it. 4. Otherwise, print a warning and set the exedir to current working directory (.). This option added in version 3.0. --expert --hfscheck --hfsdir=PATH Allows extra parameters to be passed through to storage nodes. Runs an HFS check. Specifies the root location where the /data0? data mounts are located on the storage nodes. Default value is root (/). Sets index stripe size to this number (NUM) of elements. Terminates (kills) any running Avamar processes before starting. This is not the default setting. Specifies the login manager host address. Specifies the login manager PORT number. Sets mainhost storage node to this IP address (IP-ADDR). Default value is IP address of storage node 0.0. Sets master module to N. Default value is 0.
--indexelements=NUM --kill
--masterdc=NUM
384
start.dpn COMMAND REFERENCE --matchbits=N Sets initial number of hfs index stripes. If zero is supplied as a value (N), the system dynamically computes the initial number of match bits value according to the following formula: floor(log2(ndrives)). Supplying a positive integer as a value (N), sets the initial number of hfs index stripes to that power of two. For example, --matchbits=8 sets the initial number of hfs index stripes to 256. --maxlogfiles=NUM Sets maximum number of gsan.log files to retain. Default value is 25 MB. NOTE: In multi-node servers, this setting applies equally to all storage nodes on the server (you cannot configure gsan.log file size on a node-by-node basis). --maxlogsize=MB Sets maximum gsan.log file size in megabytes (MB). Default value is 25 MB. NOTE: In multi-node servers, this setting applies equally to all storage nodes on the server (you cannot configure gsan.log file size on a node-by-node basis). --maxrwatomdatastripe=SIZE --minstripestowrite=NUM --nat --noinit Sets maximum read-write (rw) atomic data stripe SIZE to this number of elements. Sets minimum number (NUM) of stripes per family that must be online in order to write. Sets whether the server is using NATing modules. Does not run initacnt and avmaint init (pre 1.2.1 behavior). --init is the default setting. This option added in version 1.2.1. --norun | --n --nousercheck Does not actually execute commands. --run is the default setting. If supplied, normal verification that start.dpn has been invoked as user admin is bypassed. --usercheck is the default setting. PASSWORD for the --user=USER@AUTH account. When used with --init, this must be the Avamar root user account password. --probedir=PATH Specifies a PATH to a directory containing a valid probe.out file (page 477).
--password=PASSWORD
385
start.dpn COMMAND REFERENCE --quiet | -q --paritygroups=Nx,Fy Supplying either --quiet or -q causes start.dpn to run quietly. Sets parity description by specifying the sizes of near and far parity groups, where Nx is the maximum size of near parity groups and Fy is the maximum size of the first far parity group. Larger parity groups improve storage efficiency at the expense of reconstruction efficiency. In other words, large parity groups use disk space more efficiently, but take more time and more disk seeks to backup or restore data when a node or module is down. In multi-node servers, the maximum near parity setting is always the number of storage nodes in a module. Far parity groups will never grow larger than the number of modules minus one. Default value is N8,F4. --ramfsoptions=OPTIONS Controls stripes are placed in RAM, where OPTIONS is some combination of HWUG+XDCP designators, where: H W U G X C D P HFS stripes. Persistent store stripes. User accounting stripes. Delete stripes. Index stripes. Composite stripes. Data stripes. Parity stripes.
Multiple entries must be separated by commas. Default value is H+X,U+XP,G+P. This option added in version 3.6. --runlevel={admin | fullaccess} Sets server run level, to one of the following: admin Administration-only mode (only root and aroot users can access the server). Full access by all users.
fullaccess
IMPORTANT: Beginning with version 2.0.2, support for numeric run levels 4 and 6 has been discontinued in favor of using admin and fullaccess string names, respectively.
386
start.dpn COMMAND REFERENCE --rwmatchbits=N Sets initial number of read-write (rw) index stripes. If zero is supplied as a value (N), the system dynamically computes the initial number of match bits value according to the following formula: floor(log2(ndrives/4)). Supplying a positive integer as a value (N), sets the initial number of read-write (rw) index stripes to that power of two. For example, --rwmatchbits=8 sets the initial number of readwrite (rw) index stripes to 256. --savesysinfo If supplied, special storage server (GSAN) cleanliness tests are run. This is the default setting. Outputs a very abbreviated (short) report that does not contain configuration, checkpoint, garbage collect or HFS check information. Specifies SSL data PORT. Default is 29000. Refer to Stunnel (page 35) for additional information. This option added in version 3.6. --systemname=NAME Sets initial user-defined descriptive system name at server startup. NAME can be any combination of printable characters up to 32 characters in length. Passes --nodetag=MODULE.NODE parameter to storage nodes (default). Runs start.dpn as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --useram If supplied, RAM-resident stripes feature is enabled. This is not the default setting. This option added in version 4.0. --usermatchbits=N Sets initial number of user index stripes. If zero is supplied as a value (N), the system dynamically computes the initial number of match bits value according to the following formula: floor(log2(ndrives/4)) Supplying a positive integer as a value (N), sets the initial number of user index stripes to that power of two. For example, --usermatchbits=8 sets the initial number of user index stripes to 256.
--short
--sslport=PORT
--tag --user=USER@AUTH
387
start.dpn COMMAND REFERENCE --usestunnel Specifies that stunnel should be enabled when server is started. This is the default setting. Refer to Stunnel (page 35) for additional information. This option added in version 3.6. --usewritelogheaders Enables feature that periodically updates server writelog header in order to prevent writelog corruption. This is the default setting. This option added in version 3.7. --verbose | -v Supplying either --verbose or -v provides maximum information.
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. You must include the --expert option in order to use many of these advanced command-line options. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
--chunkhashcheck
Enables additional hash validation in datastripe getchunk. Requires --expert. This option added in version 1.2.1. Sets debug time-out in seconds (SEC). Requires --expert.
--dbtimeout=SEC
--dpntimecheck=SEC
If set to a positive integer, the server continually verifies that all nodes report the same time of day, within this number of seconds (SEC) tolerance. In other words, this is the maximum allowable difference between any two nodes reported time of day. Requires --expert. If set to zero (0), this feature is disabled. Completely shuts down any storage node that experiences a fatal error. Requires --expert.
--exitonfatal
388
start.dpn COMMAND REFERENCE --freezelimits="STRING" Specifies upper limits for various processor activity statistics. The statistics must stay at or below the specified limits in order to consider the server to be in an idle state. STRING must be in the following format (which is also the default setting): busy=3,diskio=40,interrupt=200,switch=10 00 busy statistic refers to the percentage of time that the CPU is busy (the opposite of being idle); diskio refers to disk reads or writes per second; interrupt refers to interrupts per second; switch refers to context switches per second. This option added in version 3.5.
Deprecated Options
Beginning with the noted release, use of these options is officially discouraged. --crc --fillindex --forcerestart --fork --inside --maxdatastripe --maxdc --maxnode --nodes --outside --poolsize --profiling --remotestart --splitcount --splitspread --squash --statnode --syncwrite IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0. IMPORTANT: Deprecated in version 4.0.
389
start.dpn COMMAND REFERENCE --useramfs IMPORTANT: Deprecated in version 4.0. Specifies whether or not to use ramfs feature. If specified, ramfs will be initialized and the server informed of its location. Default location is /ramroot. IMPORTANT: The --useramfs option is sticky. Once it has been specified, it will continue to be used after restarts or when starting new nodes. To turn off ramfs, you must use supply the --nouseramfs option. This option added in version 3.6.
Notes
User Name and If your Avamar user name and password are present in your .avamar file (page Password 426), then --id=USER@AUTH and --password=PASSWORD are not required on the
command line.
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option. It is usually sufficient to restart the Avamar server simply by running start.dpn without any parameters as long as the following conditions are satisfied: 1. A valid probe.out file (page 477) is located in /usr/local/avamar/var or in the directory specified by the SYSPROBEDIR environment variable (page 424). 2. You have run ssh-add admin_key or are prepared to directly enter the admin password as the ssh command is programmatically run on each storage node. 3. The dpnutils RPM is installed. 4. The /usr/local/avamar/bin directory is in your path.
390
start.nodes
The start.nodes program starts one or more Avamar server nodes. This should only be done when adding or replacing nodes in an Avamar server.
Synopsis
start.nodes --nodes=MODULE.NODE[,MODULE.NODE,...] [--checkpointdir=DIR] [--clean] [--copy] [--error] [--expert] [--hfsdir=PATH] [--kill] [--lmaddr=HOSTNAME] [--lmport=PORT] [--newdatacenter] [--norun] [--probedir=PATH] [--quiet] [--ramfsoptions=OPTIONS] [--runlevel={admin | fullaccess}] [--tag] [--useramfs] [--usestunnel] [--usewritelogheaders] [--verbose]
Node Descriptor
You must include the --nodes=MODULE.NODE node descriptor or the utility will not run, where MODULE.NODE is the physical node (page 478) you want to start. Multiple MODULE.NODE values can be included as a comma-separated list.
Options
--checkpointdir=DIR --clean Specifies the directory where the data is on each /data0? mount. Default value is cur. If this option is not supplied (default behavior), start.nodes checks storage nodes for any existing data and ensures that it is not overwritten (cleaned) during startup. However, if you want to restart with clean storage nodes (all existing data removed), you must supply the --clean option. --copy --error --expert --hfsdir=PATH Copies files to the storage nodes. This is the default setting. Generates error when execution fails. Allows extra parameters to be passed through to storage nodes. This is not the default setting. Specifies the root location where the /data0? data mounts are located on the storage nodes. Default value is '/'. Terminates (kills) any running Avamar processes before starting. This is not the default setting. Specifies the login manager host address. Specifies the login manager PORT number. Specifies that this new node is in a new module (datacenter) that is being created.
391
start.nodes COMMAND REFERENCE --norun --probedir=PATH --quiet --ramfsoptions=OPTIONS Does not actually execute commands. Specifies a PATH to a directory containing a valid probe.out file (page 477). Runs quietly. Controls stripes are placed in RAM, where OPTIONS is some combination of HWUG+XDCP designators, where: H W U G X C D P HFS stripes. Persistent store stripes. User accounting stripes. Delete stripes. Index stripes. Composite stripes. Data stripes. Parity stripes.
Multiple entries must be separated by commas. Default value is H+X,U+XP,G+P. This option added in version 3.6. --runlevel= {admin | fullaccess} Sets server run level, to one of the following: admin Administration-only mode (only root and aroot users can access the server). Full access by all users.
fullaccess
IMPORTANT: Beginning with version 2.0.2, all support for numeric run levels 4 and 6 has been discontinued in favor of using admin and fullaccess string names, respectively. --systemname=NAME Sets initial user-defined descriptive system name at server startup. NAME can be any combination of printable characters up to 32 characters in length. Passes --nodetag=MODULE.NODE parameter to storage nodes (default).
--tag
392
start.nodes COMMAND REFERENCE --useramfs Specifies whether or not to use ramfs feature. If specified, ramfs will be initialized and the server informed of its location. Default location is / ramroot. IMPORTANT: The --useramfs option is sticky. Once it has been specified, it will continue to be used after restarts or when starting new nodes. To turn off ramfs, you must use supply the --nouseramfs option. This option added in version 3.6. --usestunnel Specifies that stunnel should be enabled or restarted on the specified nodes. This is the default setting. Refer to Stunnel (page 35) for additional information. This option added in version 3.6. --usewritelogheaders Enables feature that periodically updates server writelog header in order to prevent writelog corruption. This is the default setting. This option added in version 3.7. --verbose Provides maximum information.
Avamar-Only Options
Avamar-only options access advanced functionality that is normally reserved for use by EMC personnel only. IMPORTANT: Misuse of these advanced options can cause loss of data. If you are unsure about any aspect of these options, contact EMC Technical Support for additional information before using them.
--exedir=PATH
Dynamically locates the gsan executable according to the following rules: 1. If --exedir is supplied, use that gsan executable without further checking. 2. If gsan executable exists in the current working directory, use it. In this case, check if the gsan executable exists in /usr/local/avamar/bin also and, if so, print a warning if ./gsan is older than /usr/local/avamar/bin/gsan. 3. If gsan executable exists in /usr/local/ avamar/bin, use it. 4. Otherwise, print a warning and set the exedir to current working directory (.). This option added in version 3.0.
393
start.nodes COMMAND REFERENCE --freezelimits="STRING" Specifies upper limits for various processor activity statistics. The statistics must stay at or below the specified limits in order to consider the server to be in an idle state. STRING must be in the following format (which is also the default setting): busy=3,diskio=40,interrupt=200,switch=10 00 busy statistic refers to the percentage of time that the CPU is busy (the opposite of being idle); diskio refers to disk reads or writes per second; interrupt refers to interrupts per second; switch refers to context switches per second. This option added in version 3.5.
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
394
stats.sh
The stats.sh shell script retrieves the following statistics: List of all client agents installed and number of each type Total number of bytes protected for all installed clients IMPORTANT: The stats.sh shell script is located in the utility node /usr/local/avamar/bin directory and must be run from that location.
395
status.dpn
The status.dpn program continuously returns status of Avamar server nodes. The status report can be sorted by node ID (default), node IP address, number of dispatchers, load average, megabytes used or percentage full. This program can also accept some avmaint options (page 53).
Synopsis
status.dpn [INTERVAL] --help --sort=[+ | -][dispatcher | full | ipaddr | load | node | used] [AVMAINT-OPTIONS]
Options
INTERVAL Time INTERVAL in seconds between status updates. Entering zero (0) returns a single status report and exits (does not loop). Sorts status report by one of the following: dispatcher full ipaddr load node used Number of dispatchers. Percentage full. Node IP address. Load average. Node ID. This is the default setting. Megabytes used.
You can also control the direction of the sort by including a plus (+) or minus (-) sign with the --sort option. Normal sorting (alphanumeric, largest values first, and so forth) is the default; reverse sorting requires the minus (-) sign. For example, supplying --sort=+used sorts the status report according to the megabytes used on each node (largest values first); supplying --sort=-node sorts the status report in reverse alphanumeric order according to the node ID. --help AVMAINT-OPTIONS Shows help, then exits. Refer to avmaint (page 53) for a complete list of options that can be included with this utility.
Notes
This program added in version 1.2.
396
store-checkpoint
The store-checkpoint program stores selected checkpoints as tar bundles. The destination Avamar server is only used for temporary storage of the checkpoints; the checkpoints cannot be used as-is on the destination server. Instead, the checkpoints must be retrieved by way of pull-checkpoint (page 334) in order to use them. The source and destination server do not have to be configured identically, but the destination server must have at least as many /data* partitions as the source server.
Synopsis
store-checkpoint [--bump=N] --cp=cp.YYYYMMDDHHMMSS[,...] [--debug] [--dryrun] --dst=NODE-LIST [--dstkey=FILE] [--gzip] [--help] --src=NODE-LIST [--srckey=FILE] [--verbose]
Options
--bump=N --cp=cp.YYYYMMDDHHMMSS[,...] Increments (bumps up) destination partition numbers by N. Specifies which checkpoint to store, where cp.YYYYMMDDHHMMSS is a valid checkpoint ID. Multiple checkpoints can be specified on the same command line; separate multiple checkpoint IDs with a comma. This option is required. --debug --dryrun --dst=NODE-LIST --dstkey=FILE --gzip --help --src=NODE-LIST --srckey=FILE --verbose Prints utility session information but does not actually perform the actions. Runs auxiliary scripts in debug mode. Specifies destination Avamar nodes (NODELIST). This option is required. Specifies path to destination Avamar server OpenSSH dpnid key FILE. Applies gzip to the tar streams. Shows help, then exits. Specifies source Avamar nodes (NODELIST). This option is required. Specifies path to source Avamar server OpenSSH dpnid key FILE. Provides maximum information.
397
Notes
IMPORTANT: You must load the dpnid OpenSSH key before running this utility. This program requires that source and destination Avamar server nodes have similar data partitioning schemes (for example, both source and destination servers have /data01 thru /data04 partitions).
Examples
Although some of these examples continue (wrap) to more than one line, all commands and options must be entered on a single command line (no line feeds or returns allowed). This example sends specified checkpoints in one tar bundle from source nodes 0.2 thru 0.6 to destination nodes 0.0 and 0.1: store-checkpoint --dst=0.0,0.1 --cp=cp.20030616080300 --cp=cp.20030616080900 --src=0.2,0.3,0.4,0.5,0.6 This example sends specified checkpoints in one tar bundle to the specified singlenode server, using that server's SSH key: store-checkpoint --dst=dpe17 --dstkey=$HOME/.ssh/dpe17-dpnid --cp=cp.20030616080300 --cp=cp.20030616080900 --src=#.#
398
stunctl
The stunctl program controls the stunnel service. The stunnel service is used in order to present a new default SSL interface to the Avamar server. Although stunctl can be used for manual control of the stunnel service, stunctl is primarily intended for use by other programs that provide control of the Avamar server such as restart.dpn (page 364), shutdown.dpn (page 378), start.dpn (page 382) and start.nodes (page 391). Refer to Stunnel (page 35) for additional information about stunnel implementation in Avamar.
Synopsis
stunctl [--debug] [--help] [--nodes=DESCRIPTOR] [--sslport=PORT] [--verbose] {help | restart | start | status | stop}
Options
--debug --help --nodes=DESCRIPTOR Prints utility session information but does not actually perform the actions. Shows help, then exits. Same as supplying the help command. Specifies which nodes should be affected by this stunctl action, where DESCRIPTOR is one or more logical node designations. Refer to Physical vs. Logical Node Numbers (page 478) for additional information. Specifies SSL accept data PORT. Default value is 29000. Provides maximum information.
--sslport=PORT --verbose
Commands
One (and only one) of the following commands must be supplied on each command line: help restart start status stop Shows help, then exits. Same as supplying the --help option. Stops, then restarts stunnel service on each node. Starts stunnel service. Show stunnel status. Stops stunnel service.
399
Examples
The following example stops stunnel services on all nodes, changes the SSL port setting to 29001, then restarts stunnel services on all nodes: stunctl restart --sslport=29001
400
suspend_crons
The suspend_crons program suspends activity of any of the Avamar server cron scripts: cp_cron (page 224), gc_cron (page 278) and hfscheck_cron (page 288). It does not affect any currently running scripts. The execution of suspended scripts can be continued with the resume_crons (page 369) script.
Synopsis
suspend_crons [--autoonly]
Options
--autoonly If supplied, automatic running of Avamar server cron scripts from dpn_crontab is turned off. If not supplied (default behavior), all invocations of Avamar server cron scripts is completely disabled. This option added in version 3.5.
Notes
The --autoonly state is only checked by cron_env_wrapper (page 228) and not the individual cron scripts (for example, cp_cron, gc_cron, morning_cron_run, and so forth). Therefore, disabling Avamar server cron scripts with --autoonly does not impact the ability to manually run Avamar server cron scripts.
401
timedist
The timedist program distributes NTP configuration files to various nodes in the Avamar server. IMPORTANT: This program must be run as user dpn. This program attempts to load the dpnid OpenSSH key from the dpn user .ssh directory. This program will fail if it is run as a different user.
IMPORTANT: Documentation for this script is provided for reference purposes only. The timedist program is typically invoked programmatically by asktime (page 43) and should not be run directly from the command line.
Synopsis
timedist [--configdir=DIR] [--debug] [--domall] [--has_utility_node] [--help] [--noactivate] [--nodes=NODE-LIST] [--nozone] [--set=DATE-TIME] [--verbose]
Options
--configdir=DIR Specifies directory (DIR) containing time configuration files created by mktime (page 318). Default location is /usr/local/avamar/var/time-config-files. Prints utility session information but does not actually perform the actions. Configures all utility nodes. Otherwise, utility nodes other than 0.m are skipped. Processes a combined 0.s/0.m node. Shows help, then exits. Does not restart ntpd. Specifies which node(s) to configure. If not supplied, all nodes are configured. Requires physical MODULE.NODE designations as defined in probe.out (page 477). Refer to Physical vs. Logical Node Numbers (page 478) for additional information. --nozone Does not not change the time zone symbolic link.
402
timedist COMMAND REFERENCE --set=DATE-TIME --verbose Sets initial date and time (DATE-TIME) value for /bin/date --set. Provides maximum information.
Notes
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
Files
Files used: ${configdir}/time-config-files/mktime.out ${configdir}/time-config-files/ntp.conf* ${configdir}/time-config-files/step-tickers* Where ${configdir} is one of the following, in order of decreasing precedence: 1. Specified by --configdir=DIR 2. ${SYSPROBEDIR}/time-config-files 3. /usr/local/avamar/var/time-config-files A complete set of time configuration files must exist, as created by a previous invocation of your custom version of mktime (page 318). Files modified on Avamar nodes: /etc/ntp.conf /etc/ntp/step-tickers /etc/localtime /etc/sysconfig/clock This application automatically rotates its log file whenever it is run and when the log file exceeds 1MB in size. Up to eight log files are retained.
403
Troubleshooting Information
The following table describes how to recover from common problems encountered when running timedist.
ERROR/SYMPTOM DESCRIPTION/REMEDY
You are using an obsolete version of timedist. Upgrade to latest version. You can also set SYSPROBEDIR, re-run probe, re-run asktime, re-run mktime.custom if asked and then re-run timedist as follows: mkdir -p /home/dpn/var export SYSPROBEDIR=/home/dpn/var probe AVAMARSERVER asktime /home/dpn/var/mktime.custom timedist Where AVAMARSERVER is the Avamar server hostname as defined in corporate DNS.
timedist: /usr/local/avamar/var/ time-config-files must exist and be a directory writable by you. When running timedist, you see any prompt for root@NODE's password.
Ensure that you correctly set the dpnid OpenSSH key before running this utility. Also verify that all nodes have the correct authorized keys (especially verify that any new nodes added to the system have the correct keys).
404
timerange
The timerange program uploads selected portions of the storage node gsan.log files by specifying a range of data/time values. It gets copied to each storage node by start.dpn (page 382). IMPORTANT: Documentation for this script is provided for reference purposes only. Do not run timerange directly from the command line.
405
timesyncmon
The timesyncmon program starts ntpd and ensure that it continues running. IMPORTANT: Documentation for this script is provided for reference purposes only. The timesyncmon program is typically invoked by timedist (page 402) or the ntpd_keepalive cron job and should not be run directly from the command line.
Synopsis
timesyncmon {--install | --keep_alive | --shutdown} [--activate] [--allow_ntpd_restart] [--archive_file=FILE.tgz] [--debug] [--help] [--local_timezone=TIMEZONE] [--no_timezone_change] [--node_type=TYPE] [--quiet] [--set=DATE_TIME] [--shutdown] [--verbose]
Commands
Commands control which operational mode (install, keep alive or shutdown) is used. One (and only one) of the following commands must be supplied on each command line: --install --keep_alive --shutdown Runs in install mode. Runs in keep-alive mode (restarts ntpd if needed). Runs in shutdown mode (stops ntpd).
Options
--activate --allow_ntpd_restart --archive_file=FILE.tgz --debug --help --local_timezone=TIMEZONE --no_timezone_change Enables and activates ntpd. Allows ntpd to be restarted if ntpd exits. Specifies name of tarball containing configuration files. Prints utility session information but does not actually perform the actions. Shows help, then exits. Specifies local TIMEZONE (for example, US/Pacific). Does not update time zone-related configuration files.
406
timesyncmon COMMAND REFERENCE --node_type=TYPE --quiet --set=DATE_TIME --verbose Used with --local_timezone to specify node TYPE (for example, s indicates utility node). Runs quietly (logging only). Specifies initial system time. Provides maximum information.
Notes
This utility runs in one of three modes: 1. Install mode (--install). This mode stops ntpd, installs time configuration files, sets the system time, starts ntpd, restarts selected processes (the ones that keep their own times), and monitors essential processes to ensure that ntpd continues to run. 2. Shutdown mode (--shutdown). This mode terminates a previously running instance of timesyncmon and stops ntpd in preparation for installing or reinstalling time configuration files and restarting ntpd. 3. Keep-alive mode (--keep_alive). This mode verifies whether ntpd is running or not, attempts to restart ntpd if it is not running and monitors essential processes to ensure that ntpd continues to run. This application automatically rotates its log file whenever it is run and when the log file exceeds 1MB in size. In install and shutdown modes, up to eight log files are retained. In Keep-alive mode, up to 24 log files are retained.
Files
The following log files are of interest on the utility node or single-node servers: /usr/local/avamar/var/timesyncmon.log /usr/local/avamar/var/cron/ntpd_keepalive_cron.log The following log files are of interest on the storage nodes in multi-node servers; /usr/local/avamar/var/timesyncmon.log /usr/local/avamar/var/ntpd_keepalive_cron.log
407
tomcatctl
The tomcatctl program controls the Apache Jakarta Tomcat service. The Tomcat service is installed along with the EMS. Although the tomcatctl command is typically invoked by dpnctl (page 239), it can be run directly from the command line.
Synopsis
tomcatctl [--debug] [--help] [--java_home=PATH] [--verbose] COMMAND
Options
--debug --help --java_home=PATH --verbose Prints utility session information but does not actually perform the actions. Shows help, then exits. Same as supplying the help command. Specifies $JAVA_HOME for final execution environment. Provides maximum information.
Commands
One (and only one) of the following commands must be supplied on each command line: help start status stop Shows help, then exits. Same as supplying the --help option. Starts Tomcat service. Shows Tomcat service status. Stops Tomcat service.
Notes
The tomcatctl program automatically loads all required OpenSSH keys. If multiple version s of Apache Jakarta Tomcat are installed, tomcatctl operates on the highest version number. Environment variables used: $JAVA_HOME $PATH Java directory (for MCS-related operations). Executable program locations.
408
truncate
The truncate program truncates the contents of a file. The hfscheck_kill (page 294) program invokes truncate in order to cause the system to re-initiate an HFS check on a checkpoint that has failed a previous HFS check attempt. This program is copied to each storage node by start.dpn (page 382). IMPORTANT: Documentation for this script is provided for reference purposes only. Do not run truncate directly from the command line.
ugcheck
The ugcheck program checks each node for upgrade-readiness. It checks each storage node that is listed in the probe.out file to verify the following upgrade requirements: CPU is 64-bit-capable. Maintenance partitions /bootalt, /rootalt and /varalt are included in /etc/fstab. Adequate disk space exists on the maintenance partitions. Sufficient free space exists on the /data01 partition to accommodate the RHEL4.6 operating system image tarballs. NOTE: By default, probe.out is located in /usr/local/avamar/var/probe.out. To select an alternate directory that contains an augmented probe.out file, set the environment variable SYSPROPEDIR. Starting with Avamar 4.1 the ugcheck program runs as a subprogram of avupos. You can also run ugcheck from the command line for a manual OS upgrade. The ugcheck program exits with a status of 0 if all nodes meet the requirements, otherwise, ugcheck exits with a non-zero status.
Synopsis
ugcheck [--help]
Options
--help Print a help message and exit
409
ugcopy
The ugcopy program distributes dpnugprep packages and files from the /data01/ugprep directory to all storage nodes. You run ugcopy from the utility node. By default, ugcopy does the following: Extracts the dpnugprep package from the latest customer tarball located in the /usr/local/avamar/src/ directory. Uses the /usr/local/avamar/var/probe.out file to identify each storage node. Then copies the dpnugprep package to the /usr/local/avamar/src/ directory on each storage node. Creates a /data01/ugprep directory on each storage node listed in the probe.out file. Then distributes the contents of the /data01/ugprep directory to the /data01/ugprep directory on each storage node. NOTE: By default, probe.out is located in /usr/local/avamar/var. To select an alternate directory that contains an augmented probe.out file, set the environment variable SYSPROPEDIR.
Synopsis
ugcopy [--(no)copy_only] [--help] [--src=DIRECTORY] [--(no)unpack]
Options
--help --(no)copy_only Prints a help message and exit. Copies files from the source directory. Default is --nocopy_only. Does not extract the dpnugprep package, and does not checksum tarballs. NOTE: The use of --copy_only implies --nounpack. --src=DIRECTORY --(no)unpack Uses DIRECTORY as the source of operating system image tarballs instead of /data01/ugprep. Does not extract the dpnugprep package Default is --unpack.
Notes
The dpnugprep RPM package includes ugcopy as part of the ugprep utility. To extract the ugcopy program from the dpnugprep package, enter the following: rpm2cpio dpnugprep-VERSION.i386.rpm | cpio -idv ./usr/local/avamar/bin/ugcopy
410
ugprep
The ugprep program is an interactive utility that upgrades Avamar nodes from Red Hat Enterprise Linux 3 (update 3, 5, or 8) to Red Hat Enterprise Linux 4 (update 6). You run ugprep from the root user on each node. This utility completes the following high-level steps: 1. Creates new ext3 filesystems on pre-existing maintenance partitions (/bootalt, /rootalt and /varalt). 2. Installs the new operating system image for Red Hat Enterprise Linux 4 on the maintenance partition. 3. Migrates a select set of configuration files from the old operating system partitions to the new partitions. 4. Updates the boot loader configuration to boot, by default, on the new operating system image.
Synopsis
ugprep [--auto] [--help] [--nocheck] [--os_src=PATH] [--verbose]
Options
--auto --help --nocheck Runs the OS upgrade without interactive prompts. Prints a help message and exit. Skips the following: Bad blocks when creating filesystems. Environment checks. Initial mount of /rootalt that is used to check existing operating system. --os_src=PATH --verbose Specifies the location of the OS image tarball when they are located in a directory other than /data01/ugprep/. Runs verbosely.
Notes
The ugprep program automatically removes /rootalt/data* mount points for which there are no corresponding entries in /etc/fstab. This helps avoid problems for tools that might expect all /data* mounts to be active mount points. The ugprep utility does not remove a mount point for /data01. If /etc/fstab does not contain a mount point for /data01, ugprep returns an error. The ugprep program writes status to the /var/log/ugprep.log file. The contents of ugprep.log are then copied to /var/log/upgrade.log for each node after ugprep finishes processing.
411
412
/rootalt/etc/pam.d/lm_* /rootalt/usr/local/avamar/etc/*
Copy of /etc/pam.d/lm_*. Copy of the contents of the /usr/local/avamar/etc/ directory, excluding node.cfg. Files in this directory include: avamar.cfg cert.pem domains.cfg dpn_crontab groups.txt install_rpm_list key.pem license.xml master_rpm_list serverlogscanners.xml stunnel/* usersettings.cfg users.txt
/rootalt/usr/local/avamar/etc /node.cfg
MCS Files The ugprep program migrates or updates the following MCS files:
/rootalt/usr/local/avamar/etc/*
Copy of the contents of the /usr/local/avamar/etc/ directory, excluding node.cfg. Examples of files in this directory relevant to MCS include dpn_crontab.
/rootalt/etc/snmp /snmpd.conf.RHEL3
/rootalt/etc/MAINT/cron/*
Copy of /etc/MAINT/cron/*, which contains Avamar cron jobs specifically related to ntpd_keepalive. Copy of the ntpd_keepalive cron job currently installed. Copy of the /etc/ntp.conf file if the file is an Avamar configuration filethe notrust tokens have been removed from the restrict statements.
413
ugprep COMMAND REFERENCE /rootalt/etc/ntp/step-tickers /rootalt/etc/sysconfig/clock /rootalt/etc/localtime /rootalt/etc/httpd/conf/ssl.crt/ server.crt /rootalt/etc/httpd/conf/ssl.key/ server.key Copy of /etc/ntp/step-tickers, which is used by ntpdate in ntpd startup. Modified copy of /etc/sysconfig/clock. File updated by using the ZONE setting in /etc/sysconfig/clock. Apache web server SSL certificate file. Apache web server key file.
/backup_utility_configs.tar
/etc/madm.conf
414
ugprep COMMAND REFERENCE /etc/MAINT/named/* Avamar DNS name server (named) configuration files. NOTE: Avamar servers no longer run DNS name servers. /etc/dhcpd.conf DHCP (dhcpd) server configuration file. NOTE: DHCP is deprecated on all Avamar servers. /var/named/* DHCP (dhcpd) server configuration file. NOTE: DHCP is deprecated on all Avamar servers. The ugprep program does not update Avamar application files on non-system partitions, such as those in the /data01 partition.The following directories are not updated: /usr/local/avamar/doc/ /usr/local/avamar/src/ /usr/local/avamar/var/ The following directories, which contain Avamar application files, remain on the root (/) partition of the old operating system: /usr/local/avamar/bin/ /usr/local/avamar/httpd/ /usr/local/avamar/httpds/ /usr/local/avamar/install/ /usr/local/avamar/lib/ /usr/local/avamar/man/
operating
Default boot label. Boots the RHEL4.6 operating system by using an SMP kernel (assuming that an SMP kernel is available for the hardware platform).
operating-up
Alternate boot label. Boots the RHEL4.6 operating system by using a uniprocessor kernel.
415
failsafe
Alternate boot label. Boots a copy of the default RHEL4.6 SMP kernel. NOTE: If using the default operating boot option fails because of a problem such as storage corruption in the /boot partition, this same type of problem might prevent you from using the failsafe boot option.
maintenance
Alternate boot label. Boots the operating system installed on the maintenance partitions, typically the previous operating system, RHEL3 (update 3, 5 or 8). NOTE: Use this boot option to run Avamar under the old operating system.
416
wait.crunch
The wait.crunch program waits for the specified type and amount of asynchronous stripe crunching to complete.
Synopsis
wait.crunch [--accounting] [--atomic] [--composite] [--help] [--hfs] [--hfsport=PORT] [--id=USER@AUTH] [--multiday] [--password=PASSWORD] [--persistent] [--probedir=PATH] [--q | --quiet] [--resetandrollover] [--rollover] [--server=AVAMARSERVER] [--timeout=SEC] [--v | --verbose]
Options
--accounting --atomic --composite --help --hfs --hfsport=PORT --id=USER@AUTH If supplied, wait applies to accounting stripes. This is not the default setting. If supplied, wait applies to atomic data stripes. This is the default setting. If supplied, wait applies to composite data stripes. This is not the default setting. Shows help, then exits. If supplied, wait applies to HFS stripes. This is the default setting. Specifies Avamar server data PORT. Authenticate on the source Avamar server as this Avamar user ID (account name). Where USER is the Avamar user name and AUTH is the authentication system used by that user. Default internal authentication domain is avamar. For example: jdoe@avamar. --multiday --password=PASSWORD --persistent --probedir=PATH --q | --quiet --resetandrollover If supplied, continue to wait after a rollover. This is not the default setting. PASSWORD for the --id=USER@AUTH account. If supplied, wait applies to persistent store stripes. This is not the default setting. Specifies a PATH to a directory containing a valid probe.out file (page 477). Runs quietly (logging only). If supplied, issue avmaint crunch resetandrollover (page 87) command before waiting.
417
wait.crunch COMMAND REFERENCE --rollover --server=AVAMARSERVER --timeout=MIN If supplied, issue avmaint crunch rollover (page 87) command before waiting. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Specifies maximum number of minutes (MIN) wait.crunch will be allowed to run. A setting of zero (0) specifies that wait.crunch be allowed to run forever (no timeout). Zero is the default setting. --v | --verbose Provides maximum information. This is the default setting.
Notes
This program added in version 4.0.
User Name and If your Avamar user name and password are present in your .avamar file (page Password 426), then --id=USER@AUTH and --password=PASSWORD are not required on the
command line.
This program requires a valid probe.out file (page 477) in order to resolve MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) typically stores the path to a directory containing a valid probe.out file. If SYSPROBEDIR is not set, the default probe.out location (/usr/local/avamar/var/) is used. You can also override this location with the --probedir=PATH option.
418
wait.dpn
The wait.dpn program waits for all the stripes to go online. It can be run on its own and it is also used by restart.dpn (page 364).
Synopsis
wait.dpn [-ap=PASSWORD] [--hfsaddr=AVAMARSERVER] [--hfsport=PORT] [--id=USER] [--offlineok] [--timeout=N]
Options
--ap=PASSWORD --hfsaddr=AVAMARSERVER --hfsport=PORT --id=USER --offlineok --timeout=N Operating system PASSWORD. AVAMARSERVER IP address or fully qualified hostname (as defined in DNS). Specify a different data PORT (only used for HFS check). Run as this operating system USER. Allow offline stripes. Set timeout to N minutes. Default value is 60.
Notes
Typically, --ap=PASSWORD, --hfsaddr=AVAMARSERVER and --id=USER values are set in a configuration file on the utility node. Therefore, these options are not normally required on the command line.
419
website
The website program performs various operations on the Avamar integrated web server, which provides the Avamar Web Access services.
Synopsis
website {create-cfg | init | restart | start | status | stop | version }
Commands
One (and only one) of the following commands must be supplied on each command line. create-cfg init restart start status stop version Creates /usr/local/avamar/etc/avamar.cfg file used by the Avamar Web Access service. Stops the web server if it is running and then performs some initialization. Stops and then starts the web server. Starts the web server if it is currently stopped. Gets the status of the web server. Stops the web server if it is currently running. Returns version of this script.
Configuring and starting Avamar Web Access services on a single-node server. Configuring and starting Avamar Web Access services on a multi-node server.
Configure
Log into the server as root. When prompted for a password, enter the root password (changeme on new servers) and press ENTER. Log into the utility node as root. When prompted for a password, enter the root password (changeme on new servers) and press ENTER.
420
3. Enter y and press ENTER. The following appears in your command shell:
--hfsaddr=10.0.254.202 --vardir=/usr/local/avamar/var --singleconn ===Done
Initialize
4. Initialize web services by entering: website init The following appears in your command shell: NOTE: Ignore any FAILED status indications during web server shutdown. This is normal.
============Initialize Webserver============= ==Shutting down website Shutting down httpd: [FAILED] ===Adding web-managers alias web-managers: support@avamar.com Detected web-managers alias ===Done
Restart
5. Restart web services by entering: website restart The following appears in your command shell: NOTE: Ignore any FAILED status indications during web server shutdown. This is normal.
============Initialize Webserver============= ==Shutting down website Shutting down httpd: [FAILED] ===Adding web-managers alias web-managers: support@avamar.com Detected web-managers alias ===Done Be sure to setup /etc/avamar/avamar.cfg (/usr/local/avamar/bin/website create-cfg) Then restart webserver with SSL enabled (/usr/local/avamar/bin/website restart)
421
website COMMAND REFERENCE 6. Verify web services by doing one of the following:
IF DO THIS
Point your web browser at http://AVAMARSERVER-IP Where AVAMARSERVER-IP is the Avamar server IP address. The Please click here to transfer to the secure login page appears.
Point your web browser at http://UTILITY-NODE-IP Where UTILITY-NODE-IP is the IP address of the Avamar server utility node. The Please click here to transfer to the secure login page appears.
422
zzdpn
The zzdpn program is a system service that implements the automated shutdown and restart feature on single-node servers. IMPORTANT: zzdpn is not intended to be run directly from the command line. In order to configure and control the automated shutdown and restart feature on your Avamar server, use dpnctl (page 239).
423
ENVIRONMENT VARIABLES
AVAMAR_INSTALL_BASEDIR_PATH
This variable is only present in the environment during Solaris client installations. It is used to set the base installation directory location. Default location is /usr/local/avamar.
AVAMAR_INSTALL_VARDIR_PATH
This variable is only present in the environment during Linux and Solaris client installations. It is used to set the var directory location. Default location is /var/avamar.
SYSPROBEDIR
Stores the path to a directory where a valid probe.out file (page 477) is located. If not set, /usr/local/avamar/var is used. Utilities that use SYSPROBEDIR are: asktime (page 43) check.dpn (page 213) mapall (page 299) mktime (page 318) nodenumbers (page 324) probe (page 328) probedump (page 330) restart.dpn (page 364) rollback.dpn (page 370) scn (page 374) ssn (page 380) start.dpn (page 382) AVAMAR 4.1 TECHNICAL ADDENDUM 424
SYSPROBEUSER
Stores an operating system user ID. If set, certain utilities run as that user. Utilities that use SYSPROBEUSER are: mapall (page 299) mktime (page 318) scn (page 374) ssn (page 380) If not set, the utility runs as user admin. Typical user IDs stored by this variable are: root, admin or dpn.
425
IMPORTANT FILES
.avamar
.avamar is a flag file. This flag file contains options that are passed to various Avamar utilities when they are invoked by this user. The .avamar file is typically found in the home directory for the admin and dpn users on utility nodes. The default .avamar file typically contains a single entry:
--flagfile=/usr/local/avamar/etc/usersettings.cfg
avagent.cfg
avagent.cfg is a flag file. This flag file is located in the Avamar client var directory. This is typically C:\Program Files\avs\var on Windows clients and /usr/local/avamar/var on Unix clients. It contains options that are passed to avagent (page 50) when it is invoked.
AVAMAR-MCS-MIB.txt
AVAMAR-MCS-MIB.txt is a plain text definition file that describes the Avamar SNMP Management Information Base (MIB). On Avamar servers, the Avamar MIB is located in /usr/local/avamar/doc/AVAMAR-MCS-MIB.txt. It also installed with Avamar Administrator in the /doc subdirectory. Refer to the Avamar System Administration Manual for additional information about monitoring an Avamar server using SNMP.
426
avscc.cfg
avscc.cfg is a flag file that contains options that are passed to avscc (page 159) when it is invoked. This flag file is located in the Avamar client var directory. avscc.cfg is typically located in C:\Program Files\avs\var on Windows clients and / usr/local/avamar/var on Unix clients.
avw_start_dpn_options.txt
avw_start_dpn_options.txt is a flag file containing server startup parameters that are read by avw_install and passed to the start.dpn command line (page 382). For more information about avw_install, refer to the Avamar Server Software Installation Manual. The only valid location for avw_start_dpn_options.txt is /usr/local/avamar/var. All valid start.dpn options are allowed in avw_start_dpn_options.txt; each option must appear as a single line of text. However, if the --nocheck option is present, it is ignored. IMPORTANT: Entries in avw_start_dpn_options.txt are not validated. Therefore, be advised that if invalid start.dpn command line options are included in the options file, then start.dpn might fail with an error.
axionfs.cfg
axionfs.cfg is a flag file that is used by axionfs (page 200) to implement the Avamar File System (AvFS) feature.
427
checkpoints.xml
checkpoints.xml stores a list of all checkpoints that have been taken by the system. Default location is /usr/local/avamar/var/checkpoints.xml. The following is a sample checkpoints.xml listing:
<?xml version ="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE checkpointlist> <checkpointlist nodecount="4"> <checkpoint tag="cp.20051215000403" isvalid="true" refcount="4" cpctime="1134605043" nodestotal="4" stripestotal="159" hfsctime="1134604408" dirstotal="4" deletable="true"> <hfscheck starttime="1134605286" nodestarttime="1134605286" nodefinishedtime="1134605302" validcheck="true" errors="0"/> <nodeidlist count="4"> <nodeidrange dcno="0" lseqno="0" useqno="3"/> </nodeidlist> </checkpoint> </checkpointlist>
428
config_info
config_info is used to configure utility nodes during factory testing and final deployment of an multi-node, single-node or Commonality Assessment Tool (CAT) system at a customer site. During factory testing, default values are used. During final deployment of an Avamar server at a customer site, config_info contains final customer network and DNS entries. The create_newconfigs utility (page 227) reads these settings and configures the utility node in that particular module accordingly. Each Avamar server configuration uses a slightly different config_info.
Multi-Node Servers
config_info files for multi-node servers contain the following entries:
ENTRY DESCRIPTION
Network hostname (as defined in DNS) for this Avamar module. Customer domain name. IP address of the network gateway. IP address of the customer DNS server. This entry can be blank. Network hostname (as defined in DNS) or IP address of the customer NTP server. IP address of the customer NTP server. This entry can be blank. Subnet and subnet mask of this module. For example, 10.0.99.0/24. IMPORTANT: The only valid network widths for these subnets are /24, /25, /26 and /27.
OTHERNETS
Subnet and subnet mask of the other Avamar module in this system. For example, 10.0.98.0/24. IMPORTANT: The only valid network widths for these subnets are /24, /25, /26 and /27.
OTHERDPNS
Network hostname (as defined in DNS) for the other Avamar module in this system (if this config_info file is used to configure the primary module, this entry will contain the DNS name of the secondary module).
429
Single-Node Servers
config_info files for single-node servers contain the following entries:
ENTRY DESCRIPTION
Network hostname (as defined in DNS) for this single-node server. Customer domain name. Subnet and subnet mask of this server. For example, 192.168.0.0/16. IP address of the network gateway. IP address of this single-node server. IP address of the customer DNS server. This entry can be blank.
Network hostname (as defined in DNS) for this single-node server. Customer domain name. Subnet and subnet mask of this server. For example, 192.168.0.0/16. IP address of the network gateway. IP address of this Avamar NDMP Accelerator. IP address of the customer DNS server. This entry can be blank.
430
CAT Servers
config_info files for CAT servers contain the following entries:
ENTRY DESCRIPTION
Network hostname (as defined in DNS) for this Avamar CAT server. Customer domain name. Subnet and subnet mask of this server. For example, 192.168.0.0/16. IP address of the network gateway. IP address of this Avamar CAT server. IP address of the customer DNS server. This entry can be blank.
Spare Nodes
config_info files for single-node servers contain the following entries:
ENTRY DESCRIPTION
Network hostname (as defined in DNS) for this spare node. Customer domain name. Subnet and subnet mask of this server. For example, 192.168.0.0/16. IP address of the network gateway. IP address of this spare node. IP address of the customer DNS server. This entry can be blank.
431
emclient.xml
emclient.xml is a Tomcat configuration file that stores Avamar Enterprise Manager configuration settings. This file is located in the EMS TOMCAT/webapps/cas/WEB-INF directory, where TOMCAT is the actual Tomcat installation base directory. Note that even though this configuration file resides on the EMS, from Tomcats perspective, Avamar Enterprise Manager is a client. emclient.xml contains the following entries:
ENTRY DESCRIPTION
Name of node running EMS. Default is localhost. Data port that will be used to communicate with EMS. Default is 8778. Scope of search. Default is session. If true, checkboxes and Select All/Clear All buttons are shown on search results pages if the number of results exceeds the defaultVisibleRows value. Default is true. Specifies maximum number of search result rows that will be shown with checkboxes and Select All/Clear All buttons. If this value is exceeded, checkboxes and Select All/Clear All buttons are not shown. Default is 25.
defaultVisibleRows
432
emserver.xml
EMS configuration file. Conforms to the preferences.dtd XML Document Type Description (DTD) referenced by the JSDK 1.4 API. This file added in version 3.5.
Document Object <root type="system"> Model (DOM) <node name="com">
<node name="avamar"> <node name="asn"> <node name="module"> <node name="datastore"> <node name="mail"> <node name="service"> <node name="mc"> <node name="ca"> <node name="compatibility"> <node name="dashboard"> <node name="dpn"> <node name="event"> <node name="mcvars">
com.avamar.asn
Global EMS settings.
ELEMENT DESCRIPTION
Data port for login servicing. Default is 8779. Data port used for EMS node communication. Default is 8781. Data port used to contact EMS to service request. Default is 8778. If set TRUE, communication with the EMS is secured using Secure Sockets Layer (SSL). If set FALSE, communication with the EMS is unsecured. Default is FALSE. This element added in version 4.0.
rmi_ssl_keystore_ap
Password for the certificate keystore in /usr/local/avamar/lib/rmi_ssl_keystore, which stores the trusted certificate used by the EMS for SSLencrypting RMI connections. This element added in version 4.0. Data port used for EMS service communication. Default is 8780. Password for the certificate keystore in /root/.keystore used by tomcat. This element added in version 4.0.
service_context_port trust_keystore_ap
433
com.avamar.asn.module.datastore
EMS database settings.
ELEMENT DESCRIPTION
Data port used to access EMS database. Default is 5556. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify.
com.avamar.asn.module.mail
EMS mail settings.
ELEMENT DESCRIPTION
admin_mail_sender_address smtpHost
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify.
com.avamar.asn.service
EMS message service.
ELEMENT DESCRIPTION
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify.
434
com.avamar.mc
EMS settings.
ELEMENT DESCRIPTION
IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.0. IMPORTANT: EMC internal use only. Do not modify. Location of Java Runtime Environment (JRE). IMPORTANT: EMC internal use only. Do not modify.
com.avamar.mc.ca
ELEMENT DESCRIPTION
ada_enabled
Shows or hides Avamar Enterprise Manager ADA menu. Default is false. IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.1.
clean_emdb_all
Number of days that EMS should retain all Avamar server monitoring information. Default is 14 days. This element added in version 4.0. Number of days that EMS should retain daily Avamar server monitoring information. Default is 2555 days (that is, approximately 7 years). This element added in version 4.0. Number of days that EMS should retain hourly Avamar server monitoring information. Default is 90 days. This element added in version 4.0. Hour of day to run clean_emdb.pl script. This element added in version 4.0. Rate in seconds for EMS server to poll all managed Avamar servers for data. Default is 600 seconds (10 minutes). IMPORTANT: Increasing this value will cause a lag in reporting managed systems status. Decreasing this value will provide closer to real time status reporting at the expense of higher system resource utilization.
clean_emdb_daily
clean_emdb_hourly
clean_emdb_start
mcs_poll_interval_seconds
435
mcs_poll_retry_interval_seconds metadata_search_enabled
IMPORTANT: EMC internal use only. Do not modify. Shows or hides Avamar Enterprise Manager Search menu. Default is false (hide Search menu). This element added in version 3.7.2. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.0. IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.0.
poll_retries tomcatDir
tomcatTarFile
com.avamar.mc.compatibility
ELEMENT DESCRIPTION
db_schema_version db_views_schema_version
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify.
com.avamar.mc.dashboard
EMS capacity dashboard settings. This entire node added in version 4.1.
ELEMENT DESCRIPTION
Defines capacity threshold, after which capacity state icon is red. Default threshold is >95%. Number of days that capacity forecast icon will be red. Default setting is <30 days. Number of days that capacity forecast icon will be yellow. Default setting is <90 days. Defines capacity threshold, after which capacity state icon is yellow. Default threshold is >80%.
436
com.avamar.mc.dpn
ELEMENT DESCRIPTION
dbmaint_path flush_cacheflag flush_delay_minutes flush_dir flush_exclude_subdirs flush_logfile flush_method flush_path flush_senddataflag flush_start_minute local_flush_dir local_flush_filename_base local_flush_retention_count
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify.
com.avamar.mc.event
EMS event service settings.
ELEMENT DESCRIPTION
eventCatalogPath
437
com.avamar.mc.mcvars
EMS program directory and file locations.
ELEMENT DESCRIPTION
binDir dataDir dbDir dumpFile jarDir libDir logDir mcDir postgresDir prefsDir sqlDir upgradeDir usrLocal varDir
IMPORTANT: EMC internal use only. Do not modify. Default is bin. IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data. IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/postgres/data. IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/mcs_data_dump.sql. IMPORTANT: EMC internal use only. Do not modify. Default is lib. IMPORTANT: EMC internal use only. Do not modify. Default is lib. IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_log. IMPORTANT: EMC internal use only. Do not modify. Default is var/mc. IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/postgres. IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/prefs. IMPORTANT: EMC internal use only. Do not modify. Default is lib/sql. IMPORTANT: EMC internal use only. Do not modify. Default is lib/upgrade. IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.0. IMPORTANT: EMC internal use only. Do not modify. Default is var.
438
/etc/hosts
/etc/hosts is an operating system configuration file that contains information regarding known hosts on the network. This provides a mechanism for resolving IP addresses to network hostnames that is independent of DNS. /etc/hosts files are found on Avamar utility nodes and are used by the system to resolve the actual IP addresses of these nodes to the official hostname and various aliases.
Where IP-ADDR is the IP address of this network host, official HOST-NAME is the fully qualified official name of this network host and ALIAS-1 thru ALIAS-n are one or more optional names for this network host. Aliases can be fully qualified (for example, host.domain.com) or unqualified (for example, host). Values are separated by any number or combination of spaces or tabs; a pound sign (#) indicates the beginning of a comment.
localhost Entry The first /etc/hosts entry must always be the localhost entry. For example:
127.0.0.1 localhost.localdomain localhost
SIngle Node These entries appear in single-node and CAT server /etc/hosts files: Servers
127.0.0.1 10.0.99.5 localhost.localdomain dpne99s.corp.com localhost dpne99s
These entries reflect default network settings used for factory tests. When deploying an Avamar server at a customer site, these entries are modified during system installation to reflect actual node network settings and hostnames in use at that customer site. For example, dpn99 is typically changed to some customer-defined Avamar server name (for example, my-dpn-01) and IP addresses are changed to reflect actual customer-defined subnets used by each Avamar module.
439
gsankeydata.xml
gsankeydata.xml is an information file generated by the gathergsankeydata utility (page 277). This file is used as an input for the AvaSpere web-based license generation mechanism. This is a typical gsankeydata.xml listing:
<?xml version ="1.0" encoding="UTF-8" standalone="yes" ?> <!DOCTYPE gsankeydatalist (View Source for full doctype...)> <gsankeydatalist customer-asset-id="A-2005001041" account-id="101010"> <gsankeydata time="1128033237" hostname="multi_axion_test4"> <macaddr name="eth0" addr="00:30:48:51:EE:B3"> <ipaddr addr="10.0.66.5" /> </macaddr> <macaddr name="eth1" addr="00:30:48:51:EE:A9" /> <sysinfo kernel-version ="Linux version 2.4.21-20.EL (bhcompile@tweety.build.redhat.com) (gcc version 3.2.3 20030502 (Red Hat Linux 3.2.3-42)) #1 Wed Aug 18 20:58:25 EDT 2004" memsize="1578557440" tzname="PDT" localtime="Thu Sep 29 15:33:57 2005" /> </gsankeydata> </gsankeydatalist>
440
gsan.log
gsan.log is the log file for the Avamar storage (gsan) processes. In a multi-node server, there is one gsan.log file present on each storage node. gsan.log is located in the /datao1/cur directory on each storage node. In a single-node server, there is only one gsan.log file. It is located in the /datao1/ cur directory.
File Size and gsan.log file size and version ing is controlled by the avmaint config (page 76) Versioning maxlogsize=MB and maxlogfiles=N parameters, respectively. This behavior can
also be controlled by supplying the --maxlogsize=MB and --maxlogfiles=N options with either start.dpn (page 382) or restart.dpn (page 364) commands. Maximum gsan.log file size is controlled by the maxlogsize=MB parameter, where MB is the maximum allowable gsan.log file size in megabytes. The default setting is 25 MB. In a multi-node server, this setting applies equally to all storage nodes on the server (you cannot configure gsan.log file size on a node-by-node basis). Once gsan.log reaches this maximum size, it is renamed to gsan.log.1. gsan.log is then emptied so that it can continue storing the most recent information. When gsan.log reaches its maximum size a second time, gsan.log.1 is renamed to gsan.log.2, gsan.log is renamed to gsan.log.1 and gsan.log is again emptied so that it can continue storing the most recent information. This mechanism continues creating new historic log files and shifting information into them until the maximum number of gsan.log files allowed on that particular storage node is reached. Once this maximum limit is reached, the oldest gsan.log file is dropped from the system. This maximum number of gsan.log files allowed on any given storage node is controlled by the maxlogfiles=N parameter, where N is the maximum number of gsan.log files allowed. The default setting is to limit the number of gsan.log files allowed on any given storage node to 100 maximum. In a multi-node server, this setting applies equally to all storage nodes on the server (you cannot configure the maximum number of gsan.log files on a node-by-node basis).
441
license.xml
license.xml is the actual license key file used by the server to determine how much storage capacity has been licensed for that particular Avamar server. license.xml is typically generated by the AvaSpere web-based license generation mechanism. It must be present in the utility node /usr/local/avamar/etc in order for the server to work. This is a typical license.xml listing:
3df6d6d4e60fa9efdc1e5a8b7ad72b231cc48d71 <?xml version ="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE licensekey> <licensekey generation-time="1130544323" account-id="101010" customer-asset-name="avamar-1" customer-asset-id="A-2005001041" expires="0" protected-data-max="80"> <macaddr addr="00:30:48:27:D1:2A"/> </licensekey>
IMPORTANT: The first line of this file contains an encrypted identifier, which was generated by by the webbased license generation mechanism when the license key file was created. This encrypted identifier must be correct and present in order for license key file to work. This file added in version 3.5.
442
/etc/avamar/domains.cfg
EMC domain configuration file. Identifies supported external authentication domains and associates a unique numerical identifier with each domain name. PAM configuration file for the login manager for domain ID # authentication domain. These files are typically symbolic links to named configuration files (for example, lm_DOMAIN_NAME). Script for activating the login manager during system startup. Contains the process ID of the currently running login manager. Login manager's LDAP configuration. Identifies the PAM module to use (/lib/security/pam_ldap.so). Login manager's NIS configuration. Identifies the PAM module to use (/lib/security/pam_unix.so). Login manager's SMB configuration. Identifies the PAM module to use (/lib/security/pam_smb.so). LDAP configuration file. Identifies the LDAP server and how to communicate with the server. This is typically a symbolic link to one of the example LDAP configuration files. Example LDAP configuration file for use with an OpenLDAP (NIS based) server. Example LDAP configuration file for use with a Windows Active Directory server. NIS client configuration file. Identifies the NIS domain and domain server. Default SMB client configuration file. Identifies the Windows domain, Primary Domain Controller (PDC) and Backup Domain Controller (BDC). Domain Name Server (DNS) client resolver configuration file. Identifies the DNS server.
/etc/pam.d/lm_#
/etc/resolve.conf
443
logs.DATE.tar
The logs.DATE.tar file is generated by the getlogs utility (page 282). It contains important log files from all nodes in the system in compressed format. The DATE portion of the filename is an eight character date code (YYYYMMDD) and six-character timestamp (HHMMSS). If you view the contents of logs.DATE.tar using the tar -tvf logs.DATE.tar command, the contents typically look like this: 0.0/nodelogs.tgz 0.1/nodelogs.tgz 0.2/nodelogs.tgz 0.m/nodelogs.tgz 0.s/nodelogs.tgz 1.0/nodelogs.tgz 1.1/nodelogs.tgz 1.2/nodelogs.tgz 1.s/nodelogs.tgz Each node in the system has a subdirectory named for its physical node number as defined in the probe.out file (page 477).
444
mccli.xml
Avamar Administrator CLI preferences file. This file contains all user-modifiable parameters for the Avamar Administrator CLI application. The factory default version of mccli.xml is located in $AVAMAR_ROOT/lib. Each time the Avamar Administrator CLI application is run, $USER_ROOT/.avamardata/var/mc/cli_data/prefs is examined to determine if a working copy of mccli.xml is present. If mccli.xml is not present in $USER_ROOT/.avamardata/var/mc/cli_data/prefs, the factory default copy of mccli.xml is copied to that location from $AVAMAR_ROOT/lib. When any Avamar Administrator CLI command is invoked, $USER_ROOT/.avamardata/var/mc/cli_data/prefs/mccli.xml will be read and those settings will be used for that command session.
com.avamar.mc.cli
ELEMENT DESCRIPTION
detail_server_stats
If true, enables detail server statistics (stats) that were previously removed. Default setting is false. This element added in version 4.1. Sets maximum number of events to retrieve. An informational message will be displayed if the limit is reached. Default is 5000. List of invalid characters for use in directory or filenames. Location of the mcclimcs.xml default options file. Default is lib/mcclimcs.xml. IMPORTANT: EMC internal use only. Do not modify.
event_monitor_display_limit
445
mcclient.xml
Avamar Administrator configuration file. Conforms to the preferences.dtd XML Document Type Description (DTD) referenced by the JSDK 1.4 API.
com.avamar.mc.gui
ELEMENT DESCRIPTION
debug detail_server_stats
Debug mode if TRUE. If true, enables detail server statistics (stats) that were previously removed. Default setting is false. This element added in version 4.1. Enables or disables check to determine whether or not the Sun Java DST update has been applied. Default is true (perform check). This element added in version 3.7.1. List of invalid characters for use in directory or filenames. Enables or disables license manager quota check. Default is true (perform quota check). Sets maximum number of directories and files that will be shown at one time when browsing backup contents. Default is 50000. This element added in version 4.1. Port number to use for communication with the MCS. Utility node (fully qualified) name or IP address.
dst_update_check
port_number service_host
446
mcclimcs.xml
The mcclimcs.xml is an XML file that stores custom mccli command parameters and profile settings that will be used when you invoke any mccli command. Default Command Parameters.The mcclimcs.xml preferences file can be used to set a default value for any mccli command parameter. Any default values set in this file are used unless another value is explicitly supplied on the command line. Additionally, these default values are global (they are used by all profiles). Profiles.Each profile is an element in the XML document and is distinguishable by the mcsprofile attribute, which identifies the name of the profile. Each profile contains a list of default options to use with the MCS specified for that profile. One profile can be designated as the default profile to use if no MCS information is specified on the command line by way of the global options. Otherwise, the profile name of the MCS is all that is required on the command line and the remainder of the options are read from the configuration file. One or all of the options can be specified on the command line to override entries in the mcclimcs.xml file. IMPORTANT: If the server hostname or data port assignment are changed for any reason (for example, after running the resite utility), or the user account name or password used to run mccli commands is changed for any reason, you must manually update the corresponding settings in your mcclimcs.xml preferences file to account for those changes.
Behavior. The factory default version of mcclimcs.xml is located in $AVAMAR_ROOT/lib. Each time the Avamar Administrator CLI application is run, $USER_ROOT/.avamardata/var/mc/cli_data/prefs is examined to determine if a working copy of mcclimcs.xml is present. If $USER_ROOT/.avamardata/var/mc/cli_data/prefs/mcclimcs.xml is not present, the factory default copy of mcclimcs.xml is copied to that location from $AVAMAR_ROOT/lib. The following is a listing of the default mcclimcs.xml default options file:
<!-- MCS Profiles --> <!-<!-<!-<!-<!-<!-default : name of default MCS profile --> mcsprofile : MCS profile name --> mcsaddr : network name or IP of MCS node --> mcsport : port to contact MCS --> mcsuserid : account on MCS --> mcspasswd : password -->
<MCSConfig default="local"> <Defaults> <Commands> <!-- Add Resource, Command, Options, Option nodes to match --> <!-- the hierarchy in the mcclisyntax.xml file --> <!-- use the Value attribute to specify the default value --> <!-<Resource Name="ResourceName"> <Command Name="CommandName"> <Options> <Option Name="OptionName" Value="OptionValue" /> </Options> </Command> </Resource>
447
Practical Examples.
<Resource Name="activity"> <Command Name="show"> <Options> <Option Name="active" Value="true" /> </Options> </Command> </Resource>
This setting constrains the mccli activity show command to only show active jobs, as if the --active=true option was supplied on the command line. Consider the following client resource settings:
<Resource Name="client"> <Command Name="add"> <Options> <Option Name="enabled" Value="true" /> <Option Name="pageport" Value="29123" /> </Options> </Command> </Resource>
These settings affect the mccli client add command so that any new client will be enabled and have its page data port set to 29123 as if the --enabled=true and --pageport=29123 options were supplied on the command line.
448
mcserver.xml
MCS configuration file. Conforms to the preferences.dtd XML Document Type Description (DTD) referenced by the JSDK 1.4 API.
Document Object <root type="system"> Model (DOM) <node name="com">
<node name="avamar"> <node name="asn"> <node name="module"> <node name="datastore"> <node name="mail"> <node name="rpt"> <node name="service"> <node name="mc"> <node name="cr"> <node name="dpn"> <node name="users"> <node name="um"> <node name="event"> <node name="lm"> <node name="mcsm"> <node name="mcvars"> <node name="mon"> <node name="rpt"> <node name="um"> <node name="wo">
com.avamar.asn
Global MCS settings.
ELEMENT DESCRIPTION
datastore_module_class mailmgr_module_class
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. This element added in version 1.2.0. Data port for client connection. Default is 7778. If set TRUE, communication with the MCS is secured using SSL. If set FALSE, communication with the MCS is unsecured. Default is FALSE. This element added in version 4.0.
port rmi_over_ssl
rmi_ssl_keystore_ap
Password for the certificate keystore in /usr/local/avamar/lib/rmi_ssl_keystore, which stores the trusted certificate used by the MCS for SSL-encrypting RMI connections. This element added in version 4.0.
449
security_module_class
IMPORTANT: EMC internal use only. Do not modify. This element added in version 1.2.0. IMPORTANT: Beginning with version 1.2, support for this element has been discontinued. System password; part of an identity that need not be authenticated by the Avamar server but is used by the CLI to log into the MCS to perform management operations.
syspasswd
sysuser
IMPORTANT: Beginning with version 1.2, support for this element has been discontinued. System user name; part of an identity that need not be authenticated by the Avamar server but is used by the CLI to log into the MCS to perform management operations.
com.avamar.asn.module.datastore
MCS database settings.
ELEMENT DESCRIPTION
activities_max_batch_size
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
activities_max_queue_size
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
activities_sleep_msec_if_max_queued
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
activities_thread_cnt
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
450
audits_max_batch_size
IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.1.
audits_max_queue_size
IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.1.
audits_sleep_msec_if_max_queued
IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.1.
audits_thread_cnt
IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.1.
centera_server
IMPORTANT: Beginning with version 3.7, support for this element has been discontinued. IMPORTANT: Beginning with version 3.5.1, support for this element has been discontinued. IMPORTANT: Beginning with version 3.5.1, support for this element has been discontinued. IMPORTANT: Beginning with version 3.5.1, support for this element has been discontinued. IMPORTANT: Beginning with version 3.5.1, support for this element has been discontinued. IMPORTANT: Beginning with version 1.2, support for this element has been discontinued in favor of the database_url element.
clients_max_batch_size
clients_max_queue_size
clients_sleep_msec_if_max_queued
clients_thread_cnt
database
451
database_port
Data port used by the MCS database. Default is 5555. This element added in version 1.2.0.
database_url
Name of database used by MCS to store operational and report data. This element added in version 1.2 and supersedes the database element.
db_schema_version _required_for_update
Minimum version of the MCS database that will support an direct update to this version of the MCS database. This element added in version 3.5.1.
dbdropviewsPath
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
dbgrantPath
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
dbviewsPath
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
ds_load_file
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
events_max_batch_size
452
events_max_queue_size
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
events_sleep_msec_if_max_queued
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
events_thread_cnt
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
ispostmasterrunning_script
IMPORTANT: Beginning with version 3.5.1, support for this element has been discontinued. Maximum number of database connections. This element added in version 1.2.0.
max_db_conn
mds_data_dir
Metadata search database data directory. This element added in version 3.5.1.
mds_database_host
mds_database_port
Metadata search database data port. Default is 5557. This element added in version 3.5.1.
mds_database_url
Metadata search database universal resource locator. This element added in version 3.5.1.
mds_log_dir
Metadata search database log file directory. This element added in version 3.5.1.
453
mds_pass
Metadata search database account password. This element added in version 3.5.1.
mds_user
Metadata search database account name. This element added in version 3.5.1.
min_db_conn
Number of MCS database connections to initially allow. Default is 10. This element added in version 3.5.1.
root_dir
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
savePGPrivilegesPath
snapup_history_max_batch_size
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
snapup_history_max_queue_size
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
snapup_history_sleep_msec_if_max_queued
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
snapup_history_thread_cnt
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.5.1.
sqlScriptPath
IMPORTANT: Beginning with version 3.5.1, support for this element has been discontinued. 454
com.avamar.asn.module.mail
Module mail settings.
ELEMENT DESCRIPTION
admin_mail_address
IMPORTANT: Beginning with version 1.1.6, support for this element has been discontinued. Sender email address for MCS mail messages. Sender email address for MCS mail messages. Default is AxionAdmin@AVAMARSERVER.DOMAIN Where AVAMARSERVER.DOMAIN is the fully qualified name of your Avamar server as defined in DNS. This element added in version 1.2.0.
admin_mail_sender_address
smtpHost
455
com.avamar.asn.service
MCS message service.
ELEMENT DESCRIPTION
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. Time in milliseconds that a worker thread waits after making a complete pass to all service queues without finding any service messages. Default is 100.
com.avamar.mc
Avamar server settings.
ELEMENT DESCRIPTION
allow_end_user_to_request_snapup
If set TRUE, users can initiate an ondemand client backup using the Avamar Windows Client Windows client system tray icon. If set FALSE, Avamar Windows Clients will not be permitted to initiate ondemand backups. Default is TRUE.
crontab_file
IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.0. External Network Address Translation (NAT) IP address. This is the IP address used by backup clients to connect to the MCS when the Avamar server is located on a private network that the client cannot access directly. If set TRUE, scheduled backups are automatically enabled immediately after MCS starts and restarts. If set FALSE, regularly-scheduled group backups must be manually enabled each time the MCS is restarted. Default is FALSE.
external_nonat_addr
group_scheduler_enabled
hfsaddr
Avamar server hostname. This name is passed to the client and is generally resolved by way of the site DNS. hfsaddr to use for MCS to Avamar server communication. Default is the same as hfsaddr. 456
local_hfsaddr
com.avamar.mc.cr
Client registration settings. This entire node added in version 1.2.
ELEMENT DESCRIPTION
allow_duplicate_client_names
Controls whether or not a client name can appear multiple places in the server domain structure. Default is FALSE. This element added in version 3.0.
client_can_init_snapups
Controls whether or not a client can initiate backups immediately following activation. Default is TRUE. This element added in version 3.0.
client_listen_port client_num_retries
Default is 28002. Number of retries before client is deemed to have stopped communicating with the MCS. Default is 2. Time in seconds before client is deemed to have stopped communicating with the MCS. Default is 10. When adding multiple clients or users with the batch client registration feature, this privilege is assigned unless another is explicitly set in the clients definition file. Default is enabled,read,backup. When adding multiple clients or users with the batch client registration feature, this domain is assigned unless another is explicitly set in the clients definition file. Default is clients. Default is 28001. IMPORTANT: EMC internal use only. Do not modify. Specifies location of location of plugin files. Default setting is /usr/local/avamar/lib/plugins. This element added in version 4.1.
client_timeout_sec
default_bulk_user_permission
default_domain
mcs_service_port plugin_catalog_dir
457
plugin_catalog_loading
Controls how which plug-in catalog is used. One of the following: always (plug-in catalog will always be reloaded from the plug-in catalog file) different (plug-in catalog will be reloaded whenever a version mismatch between the plug-in catalog file and the currently loaded plug-in catalog) newer (plug-in catalog will reload from the plug-in catalog file if the file version is newer than that already loaded) This element added in version 3.0.
plugin_catalog_xml
IMPORTANT: EMC internal use only. Do not modify. Beginning with version 4.1, this option deprecated in favor of plugin_catalog_dir.
reset_registration
com.avamar.mc.dpn
Avamar client settings for the MCS.
ELEMENT DESCRIPTION
avmaint_timeout_sec
Number of seconds to wait for an avmaint (page 53) response before killing the process. Fully qualified path and executable name of the avmaint utility (page 53). Number of seconds to wait for an avmgr (page 144) response before killing the process. Fully qualified path and executable name of the avmgr utility (page 144). Fully qualified path of the EMC software root install directory. Number of seconds to wait for an avtar response before killing the process. Fully qualified path and executable name of the avtar utility (page 171).
avmaintPath avmgr_timeout_sec
458
checkPointOverdueHours
Generate a system event if checkpoint does not occur within this number of hours. Default is 24. Optional parameters to send with an avmaint getclientmsgs command (page 97). Maximum number of entries in the activity monitor table. Default is 500. Length of time in seconds to retain completed activities. client_msgs_max_count has precedence over this setting. Default is 86400 seconds (24 hours). When executing an avmaint (page 53) or avmgr (page 144) command, this is the maximum number of successive failed attempts that will be allowed before the command will no longer be attempted. This element added in version 1.2.0.
client_msgs_maint_params
client_msgs_max_count client_msgs_max_secs
CommandRetryCount
datacenterlist_schema
IMPORTANT: EMC internal use only. Do not modify. Specifies location of internal file used to implement validation of avmaint (page 53) XML output using specific schema documents. Default setting is lib/avmaint_datacenterlist.xsd. This element added in version 3.7.
dbmaint_path
Fully qualified path of the dbmaint.sh (page 231). This element added in version 1.2.0.
diskfull
Specifies a percentage of maximum server storage capacity, that when exceeded, will generate event code 22416: The Avamar server storage has exceeded maximum + operating capacity. Default setting is 100%. This element added in version 3.0.
diskwarning
Specifies a percentage of maximum server storage capacity, that when exceeded, will generate event code 22415: The Avamar server storage is more than %d + percent full. Default setting is 80%. This element added in version 3.0. AVAMAR 4.1 TECHNICAL ADDENDUM 459
dpnCommandFailureLimit
When requesting server monitor information from the Avamar server, this is the maximum number of successive failed attempts that will be allowed before Avamar server status is set to inactive. This element added in version 1.2.0.
dpnxslfile
Relative path to the XSLT file for translating the avmaint nodelist (page 122) output to internal document format necessary for proper Avamar server status display. Optional parameters to send with an avmaint geterrors command (page 98). Number of seconds between feedcheck calls. Default setting is 60. This element added in version 3.6.
errlog_maint_params feedcheck_internal_sec
feedcheck_params
IMPORTANT: EMC internal use only. Do not modify. Parameters passed thru to feedcheck. Default settings are: --conntimeout=20 --timeout=25 --vardir=/usr/local/avamar/var This element added in version 3.6.
feedcheck_path
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.6. IMPORTANT: EMC internal use only. Do not modify. Parameters passed thru to feedstart. Default setting is --maxparallelfiles=200. This element added in version 3.6.
feedstart_params
feedstartPath
IMPORTANT: EMC internal use only. Do not modify. This element added in version 3.6. Set to --nocache if caching should not be used on flush from MCS. Root directory for the avtar (page 171) backup (flush). Comma-separated list of MCS directories that should not be included in MCS flushes. This element added in version 1.2.0.
460
flush_interval_min
Number of minutes between flush calls to avtar (required). To disable, set to zero and no automatic flushes will occur. Default is 60. Optional log file for flushes. If not specified, no log is produced. One of the following: pg-dump (dump database, then backup dump files) filesystem (backup operating system data files) Default is pg-dump. This element added in version 1.2.0.
flush_logfile flush_method
flush_path
avmgr (page 144) path where flush backups are stored (optional). If not specified, default is /MC_BACKUPS. IMPORTANT: EMC internal use only. Do not modify. Generate a system event if an MCS flushes does not occur within this number of hours. Default is 120. IMPORTANT: EMC internal use only. Do not modify. Specifies location of internal file used to implement validation of avmaint (page 53) XML output using specific schema documents. Default setting is lib/avmaint_geterrors.xsd. This element added in version 3.7.
flush_senddataflag flushOverdueMinutes
geterrors_schema
hfsCheckOverdueHours
Generate a system event if a daily system integrity check does not occur within this number of hours. Default is 48. IMPORTANT: EMC internal use only. Do not modify. Specifies directory where local MCS flushes are stored. This directory is located under the directory specified by com.avamar.mc.mcsvars.baseDir. Default is /usr/local/avamar/var/mc/server_flush. This element added in version 3.0.
local_flush_dir
461
local_flush_filename_base
IMPORTANT: EMC internal use only. Do not modify. Specifies base filename for local MCS flushes. This base filename will have the date and time appended to it. Default filename is mcflush. This element added in version 3.0.
local_flush_retention_count
Specifies maximum number of local MCS flush files that will be retained. A setting of zero (0) disables local MCS flushes. Default number of local MCS flush files to retain is 5. This element added in version 3.0.
log_results_len
lscp_interval_sec
Interval in seconds between successive avmaint lscp commands (page 119). This element added in version 1.2.0. If set TRUE, when the MCS is restarted, it will ignore any Avamar client messages sent prior to the MCS restart. Default is FALSE. This element added in version 1.2.0.
move_to_last_client_msg
move_to_last_error
If set TRUE, when the MCS is restarted, it will ignore any Avamar server error messages sent prior to the MCS restart. Default is FALSE. Interval in seconds between successive avmaint nodelist commands (page 122). Optional parameters to send with an avmaint nodelist command (page 122). IMPORTANT: EMC internal use only. Do not modify. Specifies location of internal file used to implement validation of avmaint (page 53) XML output using specific schema documents. Default setting is lib/avmaint_nodelist.xsd. This element added in version 3.7.
pollerRestartMinutes
IMPORTANT: EMC internal use only. Do not modify. This element added in version 4.0.
462
retain_checkpoints
Specifies minimum number of checkpoints to retain in the system. Avamar Administrator will not allow a user to delete a validated checkpoint if that delete operation would cause the total number of retained validated checkpoints to fall below this threshold. Default is 1. This element added in version 3.0. IMPORTANT: Beginning with version 1.2, support for this element has been discontinued. When generating the mcrollback.sh script, this value determines if it includes the password or not in the script written to hard drive (optional). If not specified, it will default to FALSE and not include the password.
rollback_include_passwd
Interval in seconds between avmaint sessions commands (page 133). Optional parameters to send with an avmaint sessions command (page 133). IMPORTANT: EMC internal use only. Do not modify. Specifies location of internal file used to implement validation of avmaint (page 53) XML output using specific schema documents. Default setting is lib/avmaint_sessions.xsd. This element added in version 3.7.
vardirFlag
Fully qualified path for var directory required by avtar (page 171).
463
com.avamar.mc.dpn.users
Avamar client user account used by the MCS.
ELEMENT DESCRIPTION
Primary Avamar administration user account. administrator user account password. User account specifically reserved for performing ondemand backups initiated from Avamar Administrator clients. backuponlyID user account password. User account that MCS should use for most activities. MCUSER user account password. User account specifically reserved for performing ondemand restores initiated from Avamar Administrator clients. restoreonly user account password. root user account password.
restoreonlyAP rootAP
com.avamar.mc.dpn.um
ELEMENT DESCRIPTION
restore_admin_can_direct_restores
If set TRUE, domain restore (read) only operators can perform directed restores.
com.avamar.mc.event
MCS event service settings.
ELEMENT DESCRIPTION
email_query_limit
Restricts the number of event descriptions sent in the custom event profile email notification. Default is 1000. If there are more events than this, the following warning will be included in the email: WARNING: Only displaying 1000 of N events. Where N is the total number of events that exist. This element added in version 1.2.0.
464
ems_audit_security_login_interval
Interval in minutes at which EMS login events are published. Default is 10 minutes. This element added in version 4.0. Location of the XML definitions for the event catalogs; the path is relative to the MCS install directory. IMPORTANT: EMC internal use only. Do not modify. This element added in version 1.2.0. IMPORTANT: EMC internal use only. Do not modify. This element added in version 1.2.0. IMPORTANT: EMC internal use only. Do not modify. This element added in version 1.2.0. IMPORTANT: EMC internal use only. Do not modify. This element added in version 1.2.0. Duration in seconds for the event throttling. This element added in version 2.0.0. Number of events that must occur before event throttling begins. This element added in version 2.0.0. Duration in seconds during which identical events must occur before event throttling begins. This element added in version 2.0.0. If set TRUE, all events are logged in pending table until processed. Mainly used for debugging to determine if events are processed properly. Default is FALSE. This element added in version 1.2.0.
eventCatalogPath
EventHandler.LogHandler
EventHandler.ProfileHandler
syslog_hostname
syslog_port
throttle_duration
throttle_repeat_count
throttle_trigger_duration
use_event_pending_table
465
com.avamar.mc.lm
License manager settings. This entire node added in version 3.5.
ELEMENT DESCRIPTION
IMPORTANT: EMC internal use only. Do not modify. IMPORTANT: EMC internal use only. Do not modify. Interval in seconds at which license expiration warnings are issued. Default is 86400 seconds (24 hours). This element added in version 4.1. Number of days until the license expires after which warnings will be issued. Default is 10 days. This element added in version 4.1. IMPORTANT: EMC internal use only. Do not modify. Enables or disables license manager quota check. Default is true (perform quota check).
license_will_expire_warn_days
lmoc
com.avamar.mc.mcsm
Management console server manager settings. This entire node added in version 3.5.
ELEMENT DESCRIPTION
capForecastDataDays
Future time period (in days) for which capacity will be forecast. Default setting is 30 days. The element added in version 4.1. Specifies minimum amount of historical data (measured in days) that must be available in order to forecast capacity usage. Default setting is 14 days. The element added in version 4.1. Specifies number of days before capReachedPercentage capacity is reached that capacity alerts should begin to be sent. Default setting is 30 days. The element added in version 4.1.
capForecastDataMinDays
capForecastReachedDays
466
capMonitorIntervalMin
Specifies how often MCS will check server capacity to determine if capReachedPercentage capacity is likely to be reached within the period specified by capForecastDataDays. Default setting is 1 day (that is, once daily). The element added in version 4.1. Specifies percentage of total server storage capacity, that when reached, will begin generating capacity alerts. Default setting is 95%. The element added in version 4.1. After first health check limit warning is acknowledged, specifies time period in which subsequent health check limit warnings will be suppressed. Default setting is 90 days. The element added in version 4.1. Specifies how often MCS will perform a server health check. Default setting is 1 day (that is, once daily). The element added in version 4.1. Specifies percentage, that when subtracted from the server read-only limit, is still considered to be acceptable capacity. Default setting is 5% (that is 95% of server read-only limit). The element added in version 4.1. Specifies how often MCS will issue capacity alerts once capReachedPercentage has been reached. Alerts will continue until the health check event has been acknowledged. Default setting is 60 minutes. The element added in version 4.1. IMPORTANT: EMC internal use only. Do not modify. Sets time in minutes after which the bytes protected valued cached by the MCS is assumed to be stale. Default is 30 minutes.
capReachedPercentage
hcDelayIntervalMin
hcMonitorIntervalMin
hcOffsetROPercentage
hcReminderIntervalMin
staleBytesProtectedMinutes
467
com.avamar.mc.mcvars
MCS program directory and file locations. This entire node added in version 2.0.
ELEMENT DESCRIPTION
baseDir
IMPORTANT: EMC internal use only. Do not modify. Default is /usr/local/avamar. Other mcsvars settings are relative (beneath) this baseDir location.
binDir
IMPORTANT: EMC internal use only. Do not modify. Default is bin. This location is relative to (beneath) the baseDir location.
dataDir
IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data. This location is relative to (beneath) the baseDir location.
dbDir
IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/postgres/data. This location is relative to (beneath) the baseDir location.
dumpFile
IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/mcs_data_dump.sql. This location is relative to (beneath) the baseDir location.
jarDir
IMPORTANT: EMC internal use only. Do not modify. Default is lib. This location is relative to (beneath) the baseDir location.
libDir
IMPORTANT: EMC internal use only. Do not modify. Default is lib. This location is relative to (beneath) the baseDir location.
logDir
IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_log. This location is relative to (beneath) the baseDir location.
mcDir
IMPORTANT: EMC internal use only. Do not modify. Default is var/mc. This location is relative to (beneath) the baseDir location.
postgresDir
IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/postgres. This location is relative to (beneath) the baseDir location.
prefsDir
IMPORTANT: EMC internal use only. Do not modify. Default is var/mc/server_data/prefs. This location is relative to (beneath) the baseDir location.
sqlDir
IMPORTANT: EMC internal use only. Do not modify. Default is lib/sql. This location is relative to (beneath) the baseDir location.
468
upgradeDir
IMPORTANT: EMC internal use only. Do not modify. Default is lib/upgrade. This location is relative to (beneath) the baseDir location.
varDir
IMPORTANT: EMC internal use only. Do not modify. Default is var. This location is relative to (beneath) the baseDir location.
com.avamar.mc.mon
MCS monitoring settings. This entire node added in version 1.2.
ELEMENT DESCRIPTION
adaFailEventIntervalMinutes
Interval in minutes at which to publish a warning event if the ADA service is not running. This element added in version 4.1. Interval in seconds at which the status of the ADA service is checked. This element added in version 4.1. Specifies command used to monitor the ADA service. This element added in version 4.1. Sets ADA monitor timeout value that determines how long to allow the adaMonitorCommand to run before terminating. This element added in version 4.1. IMPORTANT: Beginning with version 3.7, support for this element has been discontinued. IMPORTANT: Beginning with version 3.7, support for this element has been discontinued. Full path to the copyScriptToNode.pl script. Default is /usr/local/avamar/ bin/copyScriptToNode.pl. Interval in seconds at which the cp_cron cron job (page 224) will be monitored. Default is 60. Minimum interval in minutes at which the dhcp process failure event will be published. Default is 60.
adaInterval
adaMonitorCommand
adaMonitorTimeout
cent_cron_enabled
cent_cronInterval
copyScript
cp_cronInterval
dhcpdFailEventIntervalMinutes
469
dhcpdInterval
Interval in seconds at which the dhcp process will be monitored. Default is 300. Full path to the discoverServiceLocations.pl script. Default is /usr/local/avamar/bin/ discoverServiceLocations.pl. Interval in seconds at which the web restore hard drive space will be monitored. Default is 300. Full path to the monitorWebDiskSpace.pl script. Default is /usr/local/avamar/bin/ monitorWebDiskSpace.pl. Interval in seconds at which the gc_cron cron job (page 278) will be monitored. Default is 60 Interval in seconds at which the hfscheck_cron cron job (page 288) will be monitored. Default is 60. Minimum interval in minutes at which the web services failure event will be published. Default is 60. Minimum interval in minutes at which the Login Manager failure event will be published. Default is 60. Interval in seconds at which the load average on the utility node will be monitored. Default is 60. Interval in seconds at which the login manager will be monitored. Default is 60. Interval in seconds at which the metadata_cron job will be monitored. Default is 60. This element added in version 3.5.1. Full path to the monitorCron.pl script. Default is /usr/local/avamar/bin/ monitorCron.pl. Minimum interval in minutes at which the ping failure event will be published. Default is 10. Interval in seconds at which the utility node will be pinged. Default is 60. 470
discoverServiceLocationsScript
diskSpaceInterval
diskSpaceScript
gc_cronInterval
hfscheck_cronInterval
httpdFailEventIntervalMinutes
lmFailEventIntervalMinutes
loadAverageInterval
loginManagerInterval
metadata_cronInterval
monitorCronScript
pingFailEventIntervalMinutes
pingTimeInterval
postgresFailEventIntervalMinutes
Interval in minutes between publishing a database failure event if there is an error with the postgres database. Default is 60. This element added in version 2.0.0. Interval in seconds between monitoring the postgres database. Default is 60. This element added in version 2.0.0. Interval in minutes at which the MCS will attempt to query the dpn configuration to see what services are configured to run on the utility node. Default is 60. Controls whether replication operations should be monitored. If set TRUE, then daily replication operations are expected; absence of them is considered an error. If set FALSE, then daily replication operations are not expected; absence of them is not considered an error. This element added in version 3.5.1.
postgresInterval
rediscoverIntervalMinutes
repl_cron_enabled
repl_cron_monitored
Controls whether replication activities should be monitored. If set TRUE or /usr/local/avamar/etc/repl_cron.cfg exists, then replication activities will be monitored by this MCS. If set FALSE and /usr/local/avamar/etc/repl_cron.cfg does not exist, replication activities will not be monitored by this MCS. This element added in version 3.5.
repl_cronInterval
Interval in seconds at which the repl_cron job will be monitored. Default is 60. This element added in version 3.5.1. Interval in seconds at which the repl_cron cron job (page 340) will be monitored. Default is 60.
repl_cronInterval
471
sshdFailEventIntervalMinutes
Interval in minutes between publishing a ssh process failure event if there is an error with the ssh process. Default is 60. This element added in version 2.0.0. Interval in seconds between monitoring the ssh process. Default is 60. This element added in version 2.0.0. Full path to the mcs_ssh_wrapper script. Default is /usr/local/avamar/ bin/mcs_ssh_wrapper. Interval in seconds at which the web services will be monitored. Default is 120. Full path to the monitorWebServices.pl script. Default is /usr/local/avamar/ bin/monitorWebServices.pl. Minimum interval in minutes at which the ntp process failure event will be published. Default is 60. Interval in seconds at which the ntp process will be monitored. Default is 300.
sshdInterval
sshWrapperScript
webServicesInterval
webServicesScript
xntpdFailEventIntervalMinutes
xntpdInterval
com.avamar.mc.rpt
Avamar reports settings.
ELEMENT DESCRIPTION
nodelist_insert_interval_sec
com.avamar.mc.um
User manager settings. This entire node added in version 1.2.1.
ELEMENT DESCRIPTION
admin_can_direct_restores
472
com.avamar.mc.wo
MCS scheduler settings.
ELEMENT DESCRIPTION
backup_window_increment
Increment used to adjust the priority of a job using the backup window priority strategy. Breaks the job backup window into N periods. At the end of each period, the priority of the job is increased by the value of backup_window_increment. Sets extended cancel timeout value. Default is 15 minutes. This element added in version 4.0.
backup_window_num_periods
cancel_extended_window_min
cancel_normal_window_min
Sets normal cancel timeout value. Default is 5 minutes. This element added in version 4.0.
completed_job_retention_hours
Number of hours to save completed jobs in the completed queue for easy viewing in the activity monitor. Default setting is 24 hours. If set TRUE, a cache prefix is added to work order numbers. This is not the default. This element added in version 4.1.
emit_cacheprefix_for_datasets
enforce_backup_window
Boolean flag to enforce that a job cannot start or run outside of its configured backup window. Interval in seconds between building the list of jobs that are cleared to run. Number of completed jobs in the completed queue to retain for easy viewing in the activity monitor.
green_list_update_sec
max_completed_job_retention_entries
473
max_concurrent_jobs
Maximum number of concurrent jobs for the entire system. Default setting is 250. This setting is correct for a base (2x7) multi-node Avamar system. It is derived by multiplying the maximum number of jobs allowed per node (typically 20, as defined by max_jobs_per_node) by the number of storage nodes in each module (there are seven storage nodes in a base multi-node Avamar system).
max_green_list_iter
Number of times the green list is built so that a job can stay on the list without being started (if bumped off, it must wait until it is rescheduled later in the backup window). Maximum number of concurrent jobs that a single storage node can reliably handle. Default setting is 20. Number of seconds to send to the client agent to wait before checking for work. Maximum percentage of max_concurrent_jobs capacity that should actually be used by the system. Default setting is 90. For example, applying the default setting of 90% to the max_concurrent_jobs setting, yields an actual system capacity of 126 actual concurrent jobs allowed by the system.
max_jobs_per_node
no_work_response_sec
percent_of_max_concurrent_jobs
474
percent_of_max_concurrent_jobs_hfscheck
IMPORTANT: EMC internal use only. Do not modify. Maximum percentage of max_concurrent_jobs capacity that should actually be used by the system during checkpoint validations (HFS checks). This element added in version 3.0.1.
priority_aging_delay_sec
Number of seconds between incrementing a job's priority if using the priority aging scheduling strategy. Scheduling strategy that increases the priority of a job as the end of its backup window approaches. Simple scheduling strategy to increment a job's priority every priority_aging_delay_sec.
use_backup_window_priority
use_priority_aging
475
mktime.custom
Output file created by the asktime utility (page 43). This file is a customized script used to produce localized ntpd configuration files.
mktime.out
Output file created by the mktime utility (page 318).
ntp.conf*
Output file created by the mktime utility (page 318).
476
probe.out
The probe.out file is generated by the probe utility (page 328). probe verifies IP addresses assigned to each node by way of DHCP. Other Avamar utilities use probe.out to resolve simple MODULE.NODE designations into actual IP addresses. The SYSPROBEDIR environment variable (page 424) points to where probe.out is located. Default location is /usr/local/avamar/var. The probe.out file will not be correctly generated if any of the utility or storage nodes are not functioning correctly. When a probe.out file is created, it need not be changed until nodes are added or removed from the Avamar server.
Beginning on the second line, there is one line for each module in the system. Module 0 is listed on the second line, module 1 on the third line. The format of each line is:
MODULE-NAME:UTILITY_NODE_IP:NODE_0_IP,...,NODE_N_IP
Where MODULE-NAME is both the name of the module and DNS hostname of the utility node for that module; UTILITY_NODE_IP is the IP address of the utility node in the module. The remaining entries are a comma-delimited list of storage node IP addresses in that module.
477
nodenumbers output is sorted by logical node designations. Physical node numbers range from 0 to 4 because there are five storage nodes per module. Logical node numbers would initially have been the same, but as nodes were added and decommissioned new logical node numbers are added, and there are holes where nodes were decommissioned. Consider what might have happened with module 0 in the listing above to get the logical and physical node numbers listed. We started with nodes 0.0 to 0.4 with IP addresses 10.0.22.10-14, respectively. Logical node 0.2 fails, and is replaced with logical node 0.5 with IP address 10.0.22.15. With the hole created by the loss of AVAMAR 4.1 TECHNICAL ADDENDUM 478
probe.out IMPORTANT FILES logical node 0.2, logical node 0.3 becomes physical node 0.2, logical node 0.4 becomes physical node 0.3, and the new logical node 0.5 is physical node 0.4. Then logical node 0.5 fails because of a bad CPU fan, and is replaced with logical node 0.6 with IP address 10.0.22.16, which is physical node 0.4. Many programs either display node numbers, or accept node numbers as input. The following table shows the various commands with the numbering scheme each displays or accepts as input:
NUMBERING PROGRAM/UTILITY PHYSICAL LOGICAL
Avamar Administrator Server Monitor avmaint utility (page 53) check.dpn utility (page 213) gsan.log file mapall utility (page 299) nodenumbers utility (page 324) restart.dpn utility (page 364) scn utility (page 374) ssn utility (page 380) start.dpn utility (page 382) start.nodes utility (page 391)
3 3 3 3 3 3 3 3 3 3 3 3
Single-Node Server
Example probe.out listing:
NONAT avamar-1:10.0.254.82,10.0.254.82:10.0.254.82
Line 1: NONAT required; this is the only supported configuration at this time.
479
probe.out IMPORTANT FILES Line 2: Three parts, each part separated by a colon (:): Part 1: server name (avamar-1 in this example). Part 2: utility node IP address and utility node IP address (both 10.0.254.82 in this example), separated by a comma (,). Part 3: server IP address again (10.0.254.82 in this example).
Line 1: NONAT required; this is the only supported configuration at this time. Line 2: Three parts, each part separated by a colon (:): Part 1: Server name (avamar-1s in this example, often axion01 or similar). Part 2: utility node IP address and utility node IP address (10.0.5.5 in both cases), separated by a comma (,). Part 3: Storage node IP address list (10.0.5.10 thru 10.0.5.12 in this example), starting with physical node 0.0, each address separated by a comma (,).
Line 1: NONAT required; this is the only supported configuration at this time. Line 2: Three parts, each part separated by a colon (:): Part 1: Module-0 ID (avamar-1 in this example). Part 2: Module-0 utility node (0.s) IP address and utility node (0.m) IP address (10.0.5.5 in both cases). Part 3: Module 0 storage node IP address list (10.0.5.10 thru 10.0.5.12 in this example), starting with physical node 0.0, each address separated by a comma (,). Line 3: Three parts, each part separated by a colon (:): Part 1: Module-1 ID (avamar-2 in this example). Part 2: Module-1 utility node (1.s) IP address (10.0.6.5 in this example). Note that there is no MCS listed in module-1; there can only be one MCS per system. Part 3: Module-1 storage node IP address list (10.0.6.10 thru 10.0.6.12 in this example), starting with physical node 1.0, each address separated by a comma (,).
480
repl_cron.cfg
repl_cron.cfg is a flag file that is used by repl_cron (page 340) to implement replication. repl_cron.cfg is a flag file that contains various commands, options and settings that are passed to the replicate utility (page 341) by repl_cron (page 340) at run time. A sample repl_cron.cfg is provided in the /usr/local/avamar/bin directory. In order to use replication, a live customized version must be placed in /usr/local/avamar/etc. repl_cron.cfg can contain any valid replicate commands and options. Refer to replicate (page 341) for a list of valid commands and options that can be included in this flag file.
step-tickers*
Output file created by the mktime utility (page 318).
stunnel.conf
stunnel.conf stores configuration settings for the stunnel feature. It is located in the /usr/local/avamar/etc directory on each node in the system. Refer to Stunnel (page 35) for additional information.
usersettings.cfg
usersettings.cfg is a flag file. Flag files contain options that are passed to various Avamar utilities when they are invoked by this user. usersettings.cfg is typically found in the /usr/local/avamar/etc directory on utility nodes. The default usersettings.cfg file typically contains the following entries:
ENTRY DESCRIPTION
Network hostname for this Avamar server as defined in DNS. Full path to the local var directory. Full path to the local bin directory. User account exiting this utility session. Normally set to root. User account PASSWORD.
481
web.xml
web.xml is a configuration file that stores Avamar Enterprise Manager web server configuration settings. This file is located in the "/usr/local/jakarta-tomcat-5.5.9/webapps/cas/WEB-INF directory. The following preference controls how long it will take for Avamar Enterprise Manager sessions to time-out after a period inactivity:
<session-config> <session-timeout>4320</session-timeout> <!-- this is in minutes. 72 hours --> </session-config>
482
IMPORTANT DIRECTORIES
Multi-Node Server Utility Nodes
/home/admin /usr/local/avamar /usr/local/avamar/bin /usr/local/avamar/var/cron /usr/local/avamar/var/mc /var/log/messages Administrator (admin) user account home directory. Avamar base installation directory. Command-line utilities (binaries and scripts). cron job log files. MCS log files. System log files.
483
Single-Node Servers
/data01/cur/gsan.log /home/admin /usr/local/avamar /usr/local/avamar/bin /usr/local/avamar/var/cron /usr/local/avamar/var/mc /var/log/messages System log files. Administrator (admin) user account home directory. Avamar base installation directory. Command-line utilities (binaries and scripts). cron job log files. MCS log files. System log files.
Clients
/usr/local/avamar Default Avamar base installation directory (AIX, HP-UX, Linux and Solaris). Client log files. Default Avamar base installation directory (Windows). Client log files.
484
Data Types
Each column in a database view stores one of the following data types:
TYPE DESCRIPTION
bool date float8 int2 int4 int8 numeric text time timestamp varchar
Logical Boolean value (true or false). Specific calendar date (year, month, day). 8-byte floating-point number. Signed 2-byte integer (whole number). Signed 4-byte integer (whole number). Signed 8-byte integer (whole number). Exact numeric value with selectable precision. Variable-length character string. Specific time of day. Specific calendar date and time of day. Variable-length character string.
485
v_activities
This view contains a record (row) for each backup, restore or validation activity that has taken place. IMPORTANT: Beginning with version 4.0, use of this database view is deprecated in favor of v_activities_2 (page 489). All official support for this database view is likely to be discontinued in a future release.
COLUMN
TYPE
DESCRIPTION
axionsystemid bytes_excluded
varchar float8
Avamar system ID. Number of bytes intentionally excluded. Not relevant for replication activities. Not relevant for replication activities. Not relevant for replication activities. Number of bytes processed after data deduplication. Number of bytes of overhead. Not relevant for replication activities. Number of bytes processed. Not relevant for replication activities. Number of bytes unintentionally skipped (errors, and so forth). Client ID (subject to change - avoid storing externally). Client name. Client operation system. Client plugin version. Completed or terminated date. Completed or terminated time. Completed or terminated date and time. Avamar server timestamp for when backup was created. Dataset used to perform this backup (applies to group-based backups only).
bytes_modified_sent bytes_modified_not_sent bytes_new bytes_overhead bytes_scanned bytes_skipped cid client_name client_os client_ver completed_date completed_time completed_ts createtime dataset
float8 float8 float8 float8 float8 float8 varchar varchar varchar varchar date time timestamp numeric varchar
486
dataset_override
bool
True if client dataset was used instead of group dataset to perform this backup. Client domain. Expiration of backup as calculated at time of backup. Expiration of backup as calculated at time of backup. Actual dataset used in backup (applies to group-based backups only). Encryption method used. Valid values are: proprietary ssl
encryption_method
varchar
error_code error_code_summary
int4 varchar
Numeric activity status completion code. If activity did not successfully complete, a short summary of this error code. Current expiration date. Current expiration timestamp. Group name (applies to group-based backups only). Activity initiated by (applies to OnDemand Snapup only). Number of file unintentionally skipped (errors, and so forth). Not relevant for replication activities. Number of files processed. Can be zero for replication activities. Not relevant for replication activities. Name of the plugin used to perform this activity. Numeric plugin ID. Date activity was recorded. Date and time activity was recorded. Time activity was recorded.
num_of_files
float8
487
retention_policy
varchar
Retention policy used to perform this backup (applies to group-based backups only). True if client retention policy was used instead of group retention policy to perform this backup. Schedule used for scheduled backups. Recurrence interval. One of daily, weekly, yearly or monthly. Expected end date of activity. Expected end time of activity. Expected end date and time of activity. Scheduled start date. Scheduled start time. Scheduled start date and time. Unique identifier for this activity. Backup label. Is blank for replication activities. Backup number. Is blank for replication activities. Actual start date of activity. Actual start time of activity. Actual start date and time of activity. Last known status code of activity. Short summary of this status code. Type of activity. Valid values are: On-Demand Snapup Scheduled Snapup Restore Validate
retention_policy_override
bool
schedule schedule_recurrence scheduled_end_date scheduled_end_time scheduled_end_ts scheduled_start_date scheduled_start_time scheduled_start_ts session_id snapup_label snapup_number started_date started_time started_ts status_code status_code_summary type
varchar varchar date time timestamp date time timestamp varchar varchar varchar date time timestamp int4 varchar varchar
488
v_activities_2
This view contains a record (row) for each non-replication activity (that is, backup, restore or validation) that has taken place. Replication activities are stored in v_repl_actitivities (page 527).
COLUMN TYPE DESCRIPTION
axionsystemid bytes_excluded bytes_modified_sent bytes_modified_not_sent bytes_new bytes_overhead bytes_scanned bytes_skipped cid client_name client_os client_ver completed_date completed_time completed_ts createtime dataset
varchar float8 float8 float8 float8 float8 float8 float8 varchar varchar varchar varchar date time timestamp numeric varchar
Avamar system ID. Number of bytes intentionally excluded. Number of bytes modified and sent. Number of bytes modified but not sent. Number of bytes processed after data deduplication. Number of bytes of overhead. Number of bytes processed. Number of bytes unintentionally skipped (errors, and so forth). Client ID (subject to change - avoid storing externally). Client name. Client operation system. Client plugin version. Completed or terminated date. Completed or terminated time. Completed or terminated date and time. Avamar server timestamp for when backup was created. Dataset used to perform this backup (applies to group-based backups only). True if client dataset was used instead of group dataset to perform this backup. Client domain. Expiration of backup as calculated at time of backup.
dataset_override
bool
domain effective_expiration
varchar varchar
489
effective_expiration_ts effective_path
timestamp varchar
Expiration of backup as calculated at time of backup. Actual dataset used in backup (applies to group-based backups only). Encryption method used for client/ server data transfer. Choices are: proprietary ssl
encryption_method
varchar
encryption_method2
varchar
Encryption method used for client/ server data transfer. Choices are: Medium High Medium strength encryption. Strongest available encryption setting for that specific client platform. No encryption.
None
NOTE: The exact encryption technology and bit strength used for any given client-server connection is dependent on a number of factors, including the client platform and Avamar server version. Refer to your Avamar Product Security Manual for additional information. encrypt_method2_sa bool True if mcserver.xml encrypt_server_authenticate preference is set true. Otherwise, false. Numeric activity status completion code. If activity did not successfully complete, a short summary of this error code. Current expiration date. Current expiration timestamp. Group name (applies to group-based backups only). Activity initiated by (applies to OnDemand Backup only). Number of file unintentionally skipped (errors, and so forth).
error_code error_code_summary
int4 varchar
490
Number of files processed. Name of the plugin used to perform this activity. Numeric plugin ID. Date activity was recorded. Date and time activity was recorded. Time activity was recorded. Retention policy used to perform this backup (applies to group-based backups only). True if client retention policy was used instead of group retention policy to perform this backup. Schedule used for scheduled backups. Recurrence interval. One of daily, weekly, yearly or monthly. Expected end date of activity. Expected end time of activity. Expected end date and time of activity. Scheduled start date. Scheduled start time. Scheduled start date and time. Unique identifier for this activity. Backup label. Is blank for replication activities. Backup number. Is blank for replication activities. Actual start date of activity. Actual start time of activity. Actual start date and time of activity. Last known status code of activity.
retention_policy_override
bool
schedule schedule_recurrence scheduled_end_date scheduled_end_time scheduled_end_ts scheduled_start_date scheduled_start_time scheduled_start_ts session_id snapup_label snapup_number started_date started_time started_ts status_code
varchar varchar date time timestamp date time timestamp varchar varchar varchar date time timestamp int4
491
status_code_summary type
varchar varchar
Short summary of this status code. Type of activity. Valid values are: On-Demand Backup Scheduled Backup Restore Validate
v_activity_errors
This view contains a record (row) that stores the total number of times a specific event code is encountered during a specific activity.
COLUMN TYPE DESCRIPTION
Client ID. Count of the number of this event code encountered. Event code. Plugin number. Session ID.
492
v_audits
This view contains a record (row) for each audit log entry.
COLUMN TYPE DESCRIPTION
Internally generated unique ID for this audit entry. Date and time of event. Domain associated with this event. event code Values include: EM EMS END_USER MCCLI MCGUI MCS SNMP_SUB_AGENT WEB_RESTORE
role
varchar
493
object
varchar
Values include: ACTIVITY AGENT BACKUP CLIENT CP CPV CRG CRON DATASET DOMAIN EMS EVENT GC GROUP HFSCHECK MCS PLUGIN PROFILE REPL REPORT RETENTION SCHEDULE SNMP_SUB_AGENT SNMPD SYSLOGD USER ACK ACTIVATE ADD AUTH BACKUP BROWSE CANCEL COPY DELETE DISABLE EDIT ENABLE EXPORT LOGON RESTART RESUME RETIRE RUN START STOP SUSPEND VALIDATE
operation
varchar
Values include:
user_name
varchar
494
v_clients
This view contains a record (row) for each client known to the MCS. IMPORTANT: Beginning with version 4.0, use of this database view is deprecated in favor of v_clients_2 (page 499). All official support for this database view is likely to be discontinued in a future release.
COLUMN
TYPE
DESCRIPTION
agent_version allow_overtime
varchar bool
Version of agent installed. True if client can ignore scheduling window end time. See also overtime_option (page 497).
allow_userinit_snapup_file_sel allow_userinit_snapups backed_up_ts can_page checkin_ts cid client_addr client_name contact_email contact_location contact_name contact_notes contact_phone created ds_override enabled full_domain_name has_snapups
varchar varchar timestamp bool timestamp varchar varchar varchar varchar varchar varchar varchar varchar date bool bool varchar bool
Allow file selection on user initiated backups. Allow user initiated backups. Last backup date/time. True if MCS can call out to client. Last checkin date/time. Client ID (subject to change avoid storing externally). Client IP address. Client name. Contact email address. Contact location. Person to contact regarding issues with this client. Contact notes. Contact phone number. Creation date. True if client can override group dataset True if client can generate activities. Fully-qualified client location. True if the client has backups.
495
Date that client information was last modified. Client OS type. Override standard retention policy on user initiated backups.
496
overtime_option
varchar
One of the following: ALWAYS Scheduled group backups are always allowed to run past the schedule duration setting. Only the next scheduled group backup is allowed to run past the schedule duration setting. Scheduled group backups are allowed to run past the schedule duration setting until a successful backup is completed. Scheduled group backups are never allowed to run past the schedule duration setting.
NEXT
NEXT_SUCCESS
NEVER
This value is automatically set to NEXT_SUCCESS when client initially registers and is cleared after one backup successfully completes. page_addr varchar IP address used to contact this client.
497
page_port pageadr_locked plugin_for_last_backup rc_override registered registered_ts restore_only retry_cnt rp_override timeout tp_override
varchar bool varchar bool bool timestamp bool int4 bool int4 bool
Data port used to contact this client. True if address cannot be updated automatically by MCS. Plugin used for the last backup. True if client can override group retry count setting. True if client has checked into MCS. Registered date/time. True if client can only do restores. Connection retry count. True if client can override group retention policy. Connection timeout value. True if client can override group timeout period setting.
498
v_clients_2
This view contains a record (row) for each client known to the MCS.
COLUMN TYPE DESCRIPTION
agent_version allow_overtime
varchar bool
Version of agent installed. True if client can ignore scheduling window end time. See also overtime_option (page 500).
allow_userinit_snapup_file_sel allow_userinit_snapups backed_up_ts can_page checkin_ts cid client_addr client_name contact_email contact_location contact_name contact_notes contact_phone created ds_override enabled full_domain_name has_snapups modified os_type override_userinit_retpol
varchar varchar timestamp bool timestamp varchar varchar varchar varchar varchar varchar varchar varchar date bool bool varchar bool date varchar varchar
Allow file selection on user initiated backups. Allow user initiated backups. Last backup date/time. True if MCS can call out to client. Last checkin date/time. Client ID (subject to change avoid storing externally). Client IP address. Client name. Contact email address. Contact location. Person to contact regarding issues with this client. Contact notes. Contact phone number. Creation date. True if client can override group dataset True if client can generate activities. Fully-qualified client location. True if the client has backups. Date that client information was last modified. Client OS type. Override standard retention policy on user initiated backups.
499
overtime_option
varchar
One of the following: ALWAYS Scheduled group backups are always allowed to run past the schedule duration setting. Only the next scheduled group backup is allowed to run past the schedule duration setting. Scheduled group backups are allowed to run past the schedule duration setting until a successful backup is completed. Scheduled group backups are never allowed to run past the schedule duration setting.
NEXT
NEXT_SUCCESS
NEVER
This value is automatically set to NEXT_SUCCESS when client initially registers and is cleared after one backup successfully completes. page_addr varchar IP address used to contact this client.
500
page_port pageadr_locked plugin_for_last_backup rc_override registered registered_ts restore_only retry_cnt rp_override timeout tp_override
varchar bool varchar bool bool timestamp bool int4 bool int4 bool
Data port used to contact this client. True if address cannot be updated automatically by MCS. Plugin used for the last backup. True if client can override group retry count setting. True if client has checked into MCS. Registered date/time. True if client can only do restores. Connection retry count. True if client can override group retention policy. Connection timeout value. True if client can override group timeout period setting.
501
v_compatibility
This view stores MCS database compatibility information.
COLUMN TYPE DESCRIPTION
component
varchar
version
varchar
502
v_datasets
This view contains a record (row) for each dataset known to the MCS.
COLUMN TYPE DESCRIPTION
True if dataset will save all data. Avamar domain associated with dataset. True if dataset is pointer to another dataset. True if dataset cannot be modified. Name of dataset if is_link is true. Name of dataset.
503
v_dpnsummary
This view contains a record (row) for each backup, restore or validation activity on a client-by-client basis.
COLUMN TYPE DESCRIPTION
clientver host mod_sent modnotsent numfiles nummodfiles operation os overhead pcntcommon reduced root seconds sessionid starttime
varchar varchar float8 float8 float8 float8 varchar varchar float8 int4 float8 varchar float8 varchar timestamp
Version of agent software running on client. Client name. Bytes modified and sent. Bytes modified but not sent. Number of files processed. Number of files modified. Type of activity reported by this entry. Client operating system. Number of bytes of overhead sent and stored on storage subsystem. Percent data deduplication. Bytes reduced by compression. Dataset used in backup applies to group-based backups only. Completed or terminated date and time. Unique identifier for client to storage subsystem session for this activity. Date and time job was actually dispatched to client by Avamar Administrator. NOTE: Start time in client log might be slightly later due to communication and job setup latency.
startvalue
float8
Scheduled start date and time, expressed as elapsed time (in seconds) since beginning of the Unix epoch. Success/fail result of this activity. Number of bytes processed. Unique identifier for workorder for this activity.
504
v_dpn_stats
This view contains a record (row) for Avamar server statistics.
COLUMN TYPE DESCRIPTION
Megabytes protected. Date. Date and time. Avamar server name. Time.
v_ds_commands
This view contains a record (row) for each optional plugin command defined for each dataset.
COLUMN TYPE DESCRIPTION
Name of command. Name of dataset. Domain. Name of plugin. Type of optional plugin command. Value associated with command name.
v_ds_excludes
This view contains a record (row) for each exclude definition defined for each dataset.
COLUMN TYPE DESCRIPTION
505
v_ds_includes
This view contains a record (row) for each include definition defined for each dataset.
COLUMN TYPE DESCRIPTION
v_ds_targets
This view contains a record (row) for each source target defined for each dataset.
COLUMN TYPE DESCRIPTION
506
v_ev_catalog
This view contains a record (row) for each event code in the events catalog.
COLUMN TYPE DESCRIPTION
Event category. Event code. Event name. Values include: ACTIVITY AGENT BACKUP CLIENT CP CPV CRG CRON DATASET DOMAIN EMS EVENT GC GROUP HFSCHECK MCS PLUGIN PROFILE REPL REPORT RETENTION SCHEDULE SNMP_SUB_AGENT SNMPD SYSLOGD USER
507
operation
varchar
Values include: ACK ACTIVATE ADD AUTH BACKUP BROWSE CANCEL COPY, DELETE DISABLE EDIT ENABLE EXPORT LOGON RESTART RESUME RETIRE RUN START STOP SUSPEND VALIDATE
Event severity. Single-line event description. Software modules generating event. Event type.
508
v_ev_cus_body
This view contains a record (row) listing the attachments for each custom events profile.
COLUMN TYPE DESCRIPTION
attachments epid
varchar varchar
v_ev_cus_cc_list
This view contains a record (row) listing the email cc recipients for each custom events profile.
COLUMN TYPE DESCRIPTION
cc_list epid
varchar varchar
v_ev_cus_codes
This view contains a record (row) for each event code that should be included in a custom events profile.
COLUMN TYPE DESCRIPTION
Event code to monitor. True if code triggers email and syslog notification. Original default setting for email and syslog notification. Unique ID for this events profile.
509
v_ev_cus_prof
This view contains a record (row) for each custom events profile.
COLUMN TYPE DESCRIPTION
True if profile is enabled. Profile domain. True if email notification should occur. Unique ID for this events profile. True if logs are included in email. True if nodelist is included in email. True if email attacments are included in the body of the email. Date and time of last email check, expressed as elapsed time (in seconds) since beginning of the Unix epoch. Directory location of log files. Name of custom profile. True if not modifiable. Email schedule. True if snmp notification should. Email subject header string. True if syslog notification should occur.
510
v_ev_cus_prof_params
This view contains event code-specific parameters for custom event profiles.
COLUMN TYPE DESCRIPTION
v_ev_cus_rpt
This view contains a record (row) for each report emailed with an event profile.
COLUMN TYPE DESCRIPTION
True if option to email report was set. Profile ID. True if report was emailed in Comma-Separated Values (CSV) format. True if report was emailed in plain text format. True if report was emailed in XML format. Report ID. Number of days, weeks or months since last email was sent, or 0 if since_option is last_notified. Unit of measure for since_count. Values include: day week month last_notified
511
v_ev_cus_snmp_contact
This view contains a record (row) for each profiles snmp trap configuration.
COLUMN TYPE DESCRIPTION
Name of the snmp community to send traps as. Profile ID. Host to send snmp traps to. Data port to send snmp traps to (default: 162).
v_ev_cus_syslog_contact
This view contains a record (row) for each custom event profile that uses syslog as the notification mechanism.
COLUMN TYPE DESCRIPTION
epid facility
varchar int4
Unique ID for this events profile. Syslog facility. Valid values are: 1 16 17 18 19 20 21 22 23 user. local0. local1. local2. local3. local4. local5. local6. local7.
format
int4
syslog_host syslog_port
varchar int4
512
v_ev_cus_to_list
This view contains a record (row) listing the email recipients for each custom events profile.
COLUMN TYPE DESCRIPTION
epid to_list
varchar varchar
513
v_ev_unack
This view contains a record (row) for each unacknowledged event logged by the MCS.
COLUMN TYPE DESCRIPTION
audience category
varchar varchar
Intended audience of event. Event category. Valid values are: APPLICATION SECURITY SYSTEM USER
Event code. Event data. Date of event. Long event description. Domain associated with the event. Internally-generated event ID. Event notes text. Event remedy text. Event severity. Valid values are: NODE NODE_FATAL OK PROCESS PROCESS_FATAL SYSTEM_FATAL USER USER_FATAL
Software modules generating event. Host generating event. Single-line event description. Time of event. Date and time of event, expressed as elapsed time (in seconds) since beginning of the Unix epoch. Event type. Valid values are: INTERNAL ERROR WARNING INFORMATION DEBUG
type
varchar
514
v_events
This view contains a record (row) for each event logged by the MCS.
COLUMN TYPE DESCRIPTION
audience category
varchar varchar
Intended audience of event. Event category. Valid values are: APPLICATION SECURITY SYSTEM USER
Event code. Event data. Date of event. Long event description. Domain associated with the event. Internally-generated event ID. Event notes text. Event remedy text. Event severity. Valid values are: NODE NODE_FATAL OK PROCESS PROCESS_FATAL SYSTEM_FATAL USER USER_FATAL
Software modules generating event. Host generating event. Single-line event description. Time of event. Date and time of event, expressed as elapsed time (in seconds) since beginning of the Unix epoch. Event type. Valid values are: INTERNAL ERROR WARNING INFORMATION DEBUG
type
varchar
515
v_gcstatus
This view contains a record (row) for each garbage collect (GC) operation.
COLUMN TYPE DESCRIPTION
bytes_recovered chunks_deleted elapsed_time end_time gcstatusid indexstripes_processed indexstripes_total node_count result start_time
int8 int4 int8 timestamp int8 int4 int4 int4 varchar timestamp
Number of bytes recovered in this garbage collect operation. Number of chunks deleted in this garbage collect operation. Total elapsed time in seconds for this garbage collect operation. Date and time this garbage collect operation ended. Unique ID for this garbage collect operation. Number of index stripes involved in this garbage collect operation. Number of index stripes. Number of nodes involved in this garbage collect operation. String result code. Date and time this garbage collect operation started.
516
v_group_members
This view contains a record (row) for each client organized by group assignment. NOTE: group. One client can be a member of more than one
COLUMN
TYPE
DESCRIPTION
cid client_name dataset_name enabled group_name restore_only retention_name use_client_ds use_client_retry use_client_rp use_client_timeout
varchar varchar varchar bool varchar bool varchar bool bool bool bool
Client ID (subject to change - avoid storing externally). Client name. Dataset name. True if client enabled in the group. Group name. True if client has been deleted and is available only for restore. Retention policy name. True if client dataset should be used. True if client retry should be used. True if client retention policy should be used. True if client timeout should be used.
517
v_groups
This view contains a record (row) for each group known to the MCS.
COLUMN TYPE DESCRIPTION
created dataset_name domain enabled failed_stop modified name priority read_only retention_name retry_cnt run_once schedule_name skip_next target_dpn timeout_min
date varchar varchar bool bool date varchar int4 bool varchar int4 bool varchar bool varchar int4
Creation date. Dataset name. Domain. True if group is active/enabled. True if group backups should stop on failed backup. Last modified date. Group name. Group priority. True if group cannot be modified. Retention policy name. Retry count. True if running only one backup. Schedule name. True if skipping next scheduled backup. Avamar server to be used for this group. Timeout in minutes.
518
v_node_space
This view contains a record (row) of disk capacity data retrieved or calculated per disk and per node.
COLUMN TYPE DESCRIPTION
capacity_mb date date_time disk diskreadonly node stripes_reserved_mb stripes_used_mb time used_mb utilization
float8 date timestamp int2 int2 varchar float8 float8 time float8 numeric
Disk size. Date. Date and time. Disk number. Value applied to normalize percent full. Node number. Bytes reserved for stripe usage. Amount of reserved stripe bytes used. Time. Disk capacity used. Percentage of storage space used.
519
v_node_util
This view contains a record (row) of node statistics retrieved or calculated per node at a particular date and time.
COLUMN TYPE DESCRIPTION
cpu_sys_percentage cpu_user_percentage date date_time disk_reads_per_sec disk_writes_per_sec diskreadonly load_avg net_in_kbytes_per_sec net_out_kbytes_per_sec net_ping node state time utilization
numeric numeric date timestamp int4 int4 int2 numeric int4 int4 numeric varchar varchar time numeric
Percentage of node utilization by operating system. Percentage of node utilization by user. Date. Date and time. Disk reads per second. Disk writes per second. Value applied to normalize percent full. Load average. Network received (KB/s). Network transmitted (KB/s). Node ping time. Node ID. Node state. Time. Percentage of storage space used.
520
v_plugin_can_restore
This view contains a record (row) of allowable plugin substitutions for restores. Each record is a one-to-one allowable substition in which the orignal backup plugin (build, version) is matched with with an allowable substitute plugin ID (can_restore_pid).
COLUMN TYPE DESCRIPTION
An exception to the plugin version value if not ALL. PID of the plugin which this plugin can use to perform restores. Numeric plugin ID. Plugin version.
v_plugin_catalog
This view contains a record (row) for each known plugin.
COLUMN TYPE DESCRIPTION
Content description of the plugin. Descriptive name of the plugin. Encryption method used. Valid values are: proprietary ssl
explicit_target_supported
bool
True if targets for the plugin can be entered when creating/editing a dataset for the plugin. True if the concept of all systems for the plugin is supported when creating/editing a dataset. True if implicit target included by default when creating/editing a dataset. True if multiple targets can be entered when creating/editing a dataset for the plugin. Name of the plugin. Unique plugin identification. Plugin version.
implicit_target_supported
bool
include_implicit_as_default
bool
multiple_targets_supported
bool
521
v_plugin_depends_upon
This view contains a record (row) for each known plugin dependency. Each record is a one-to-one match of an existing plugin ID (build, version) and the plugin ID on which it is dependent (dependence_on_pid).
COLUMN TYPE DESCRIPTION
An exception to the plugin version value if not ALL. PID of the plugin that this plugin depends upon. Numeric plugin ID. Plugin version.
v_plugin_flag_groups
This view contains a record (row) for each grouping of plugin options (flags).
COLUMN TYPE DESCRIPTION
Order of group. Text shown when cursor hovers over plugin. One of the following: logical radio
522
v_plugin_flag_pulldown
This view contains a record (row) for each entry in a pulldown plugin option (flag).
COLUMN TYPE DESCRIPTION
An exception to the plugin version value if not ALL. Displayable value of the entry. Entry in the pulldown menu. Flag ID. Order of flag in pulldown. Numeric plugin ID. Plugin version.
523
v_plugin_flags
This view contains a record (row) for each plugin option (flag) available to be used for backups and restores for each plugin.
COLUMN TYPE DESCRIPTION
An exception to the plugin version value if not ALL. Control grouping. One of the following: restore backup
description fid flag_order max min name plugin_number pidnum tooltip type
varchar varchar int4 int4 int4 varchar int4 int4 varchar varchar Numeric plugin ID. Plugin number that this flag should be directed too. Text shown when cursor hovers over plugin. One of the following: boolean (check box) integer (text box) string (text box) Flag ID. Order of group. Maximum value of the flag, if applicable. Minimum value of the flag, if applicable.
value version
varchar varchar
524
v_plugin_options
This view contains a record (row) of available options for a plugin; one option per record (row).
COLUMN TYPE DESCRIPTION
An exception to the version if not ALL. For disable options only. True if the option value is preserved on upgrades. Valid values are: browse_supported disable_browse disable_mc_adhoc_snapups disable_restore disable_validate disable_scc_adhoc_snapups disable_scheduled_snapups restore_supported snapup_supported snapup_supports_cl_options snapup_supports_exclusion snapup_supports_inclusion validate_supports
v_plugin_state
This view contains a record (row) that stores the state of each plugin.
COLUMN TYPE DESCRIPTION
An exception to the plugin version values if not ALL. True if the plugin is obsoleted. Comment as to why the plugin was obsoleted. Numeric plugin ID. True if the user added the build. Plugin version.
525
v_plugins
This view contains a record (row) for each plugin installed on any client known to the MCS.
COLUMN TYPE DESCRIPTION
Last backup date using this plugin. Plugin build. Client ID (subject to change - avoid storing externally). Name of client. Date this plugin type was first registered with MCS. Date this current plugin version was first registered with MCS. Description of plugin. Name of plugin. Plugin version.
526
v_repl_activities
This view contains a record (row) for each replication activity that has taken place.
COLUMN TYPE DESCRIPTION
bytes_excluded bytes_modified_sent bytes_modified_not_sent bytes_new bytes_overhead bytes_reduced_comp bytes_scanned bytes_skipped cid client_name client_os client_ver completed_ts dpn_domain encryp_method
float8 float8 float8 float8 float8 float8 float8 float8 varchar varchar varchar varchar timestamp varchar text
Number of bytes intentionally excluded. Number of bytes modified and sent. Number of bytes modified but not sent. Number of bytes processed after data deduplication. Number of bytes of overhead. Number of bytes reduced by compression. Number of bytes processed. Number of bytes unintentionally skipped (errors, and so forth). Client ID (subject to change - avoid storing externally). Client name. Client operation system. Client plugin version. Date and time this replication operation actually ended. Client domain. Encryption method used for client/ server data transfer. Choices are: proprietary ssl NOTE: This column is deprecated and exists for historical purposes only. Use encryp_method2 instead.
527
encryp_method2
varchar
Encryption method used for client/ server data transfer. Choices are: High Strongest available encryption setting for that specific client platform. Medium strength encryption. No encryption.
Medium None
NOTE: The exact encryption technology and bit strength used for any given client-server connection is dependent on a number of factors, including the client platform and Avamar server version. Refer to your Avamar Product Security Manual for additional information. encrypt_method2_sa bool True if server authentication was enforced at the time of the backup (that is, mcserver.xml encrypt_server_authenticate preference is set true). Numeric activity status completion code. Last known error code summary. Activity initiated by this user or MCS. Number of file unintentionally skipped (errors, and so forth). Number of files modified. Number of files processed. Can be zero for replication activities. Plugin name. Plugin number. Date replication occurred.
528
retention_type
varchar
This replicatication activity included one or more of the following retention types: D W M Y N Daily backups. Weekly backups. Monthly backups. Yearly backups. Backups not tagged as having a specific retention type.
Date and time this replication operation was scheduled to end. Date and time this replication operation was scheduled to occur. Unique identifier for this activity. Date and time this replication operation actually started. Numeric status code. Status code summary. Avamar system ID. Type of activity. Valid values are: Replication Destination Replication Source
wid
varchar
529
v_report_filter
This view contains a record (row) for each report identifying its filter options.
COLUMN TYPE DESCRIPTION
v_reports
This view contains a record (row) for each report.
COLUMN TYPE DESCRIPTION
True if a query statement is being used instead of filtering options. Report domain. Not currently supported. Report name. True if the report can not be edited or deleted. Used for 'canned' reports. Report ID. SQL statement if adhoc_query is true. Database view used by this report.
530
v_retention_policies
This view contains a record (row) for each retention policy known to the MCS.
COLUMN TYPE DESCRIPTION
daily domain duration enabled expiration_date is_link link_name monthly name override policy_no
int4 varchar numeric bool numeric bool varchar int4 varchar bool int4
Advanced policy daily retention. Domain. Duration of retention. True if enabled. Expiration date. True if this is a reference to another retention policy. Name of retention policy if is_link is true. Advanced policy monthly retention. Name of retention policy. True if advanced policy used for scheduled backups. Policy number. Valid policy numbers are: 0 1 2 3 Undefined. Compute expiration date. Static expiration date. No expiration date.
read_only unit
bool int4
True if retention policy cannot be modified. Duration unit. Valid duration units are: 0 1 2 3 4 No expiration Days Weeks Months Years
weekly yearly
int4 int4
531
v_sch_recurrence
This view contains a record (row) for each recurring schedule known to the MCS.
COLUMN TYPE DESCRIPTION
modifier
text
Qualifies entries in the value column as follows: day Indicates that this is a monthly schedule that runs on every numerical calendar day specified by the value column entry. Indicates that this is a weekly schedule that runs on every day of the week specified by the value column entry. Indicates that this is a monthly schedule that runs during the first week of the month on the day of the weekl specified by the value column entry. Indicates that this is a monthly schedule that runs during the fourth week of the month on the day of the weekl specified by the value column entry. Indicates that this is a daily schedule that runs on every hour of the day specified by the value column entry. Indicates that this is a monthly schedule that runs during the last week of the month on the day of the weekl specified by the value column entry. Indicates that this is a monthly schedule that runs during the second week of the month on the day of the weekl specified by the value column entry. Indicates that this is a monthly schedule that runs during the third week of the month on the day of the weekl specified by the value column entry.
every
first
fourth
hour
last
second
third
name
varchar
532
recur_interval
text
Recurrence interval. Valid recurrence intervals are: DAILY HOURLY WEEKLY MONTHLY YEARLY
value
text
Recurrence value. For DAILY schedules, this value is the hour of the day. For WEEKLY schedules, this value is the day of week, such as Saturday, Sunday, Monday, Tuesday, Wednesday, Thursday, and Friday. For MONTHLY schedules that repeat on a specific day of the month, this numerical value is the day of the month. For MONTHLY schedules that repeat on a specific day of a specific week, this value is the day of week, such as Saturday, Sunday, Monday, Tuesday, Wednesday, Thursday, and Friday.
533
v_schedules
This view contains a record (row) for each schedule known to the MCS. IMPORTANT: Beginning with version 4.0, use of this database view is deprecated in favor of v_schedules_2 (page 536). All official support for this database view is likely to be discontinued in a future release.
COLUMN
TYPE
DESCRIPTION
Schedule description. Domain. True if schedule is enabled/active. Type of schedule termination setting. Valid values are: 2 3 4 Never end. Run N number of times. End on a specific date.
end_recur
numeric
End recurrence. This is a specific date or a count of the number of times the schedule should run or 0 if the schedule never ends. This value is related to the value of end_policy. First start. True if this is a reference to another schedule. Last check. Last started. Schedule name if is_link is true. Minimum interval. Name of schedule. True if schedule end time can be overridden. True if cannot be modified. Recurrence counter. Recurrence interval. Valid recurrence intervals are: DAILY HOURLY WEEKLY MONTHLY YEARLY
first_start is_link last_check last_start link_name min_interval name overtime read_only recur_counter recur_interval
timestamp bool timestamp timestamp varchar timestamp varchar bool bool numeric varchar
534
Duration of start scheduling window. Start time for scheduling window. Time zone of where the schedule was created or last modified. Total duration of scheduling window. Type of schedule. The only valid schedule type is CALENDAR.
535
v_schedules_2
This view contains a record (row) for each schedule known to the MCS.
COLUMN TYPE DESCRIPTION
Schedule description. Domain. True if schedule is enabled/active. Type of schedule termination setting. Valid values are: 2 3 4 Never end. Run N number of times. End on a specific date.
end_recur
numeric
End recurrence. This is a specific date or a count of the number of times the schedule should run or 0 if the schedule never ends. This value is related to the value of end_policy. First start. True if this is a reference to another schedule. Last check. Last started. Schedule name if is_link is true. Minimum interval. Name of schedule. True if schedule end time can be overridden. True if cannot be modified. Recurrence counter. Recurrence interval. Valid recurrence intervals are: DAILY WEEKLY MONTHLY ADHOC
first_start is_link last_check last_start link_name min_interval name overtime read_only recur_counter recur_interval
timestamp bool timestamp timestamp varchar timestamp varchar bool bool numeric varchar
Duration of start scheduling window. Start time for scheduling window. Time zone of where the schedule was created or last modified.
536
total_duration type_enum
timestamp varchar
Total duration of scheduling window. Type of schedule. The only valid schedule type is CALENDAR.
v_systems
This view contains a record (row) for each Avamar system.
COLUMN TYPE DESCRIPTION
Avamar server ID. User assigned name. IP address of server. Port address of server. Last updated timestamp. Local IP address of server. Port address of MCS for server from axion_systems. Numeric system ID (automatically assigned by MCS).
537