Вы находитесь на странице: 1из 6

VCS CheatSheet 0). Daemons to look for: had hashadow 1).

Checking the status of the cluster: hastatus sum -- SYSTEM STATE -- System State A csdev94 A csdev95 RUNNING RUNNING Probed Y Y Y Y Y Y Y Frozen 0 0 AutoDisabled N N N N N N N State

-- GROUP STATE -- Group System B B B B B B B InfxSG InfxSG OracleSG OracleSG aopem01 db2SG db2SG csdev94 csdev95 csdev94 csdev95 csdev94 csdev94 csdev95

ONLINE OFFLINE ONLINE OFFLINE OFFLINE OFFLINE ONLINE

2). Status of LLT root@csdev94: /sbin/lltconfig LLT is running If LLT is not running execute: lltconfig -c

root@csdev94: /sbin/lltconfig -a list Link 0 (eri0): Node 0 csdev94 : 00:03:BA:44:B2:EC permanent Node 1 csdev95 : 00:03:BA:4D:59:D0 Link 1 (qfe3): Node 0 csdev94 Node 1 csdev95 : 00:03:BA:44:B2:EC permanent : 00:03:BA:4D:59:D0

root@csdev94: /sbin/lltstat -nvv | more

LLT node information: Node State Link Status Address * 0 csdev94 OPEN eri0 UP 00:03:BA:44:B2:EC qfe3 UP 00:03:BA:44:B2:EC 1 csdev95 OPEN eri0 UP 00:03:BA:4D:59:D0 qfe3 UP 00:03:BA:4D:59:D0 2 CONNWAIT eri0 DOWN qfe3 DOWN Should see 2 interfaces up on each side of the cluster. 3). Checking the status of GAB root@csdev94: gabconfig -a GAB Port Memberships =============================================================== Port a gen a0052c19 membership 01 -> GAB is working Port h gen 30dbfb18 membership 01 -> HAD is working If GAB is not listed start gab by executing the command found in /etc/gabtab If HAD is not listed ps aef | grep ha -> look for had and hashadow

If had or hashadow are not running execute /etc/rc3.d/S99vcs start 4). Logfiles to monitor: /var/VRTSvcs/log engine_A.log - main log 5). Other commands: root@csdev94: hagrp -list InfxSG csdev94 InfxSG csdev95 OracleSG csdev94 OracleSG csdev95 aopem01 csdev94 aopem01 csdev95 db2SG csdev94

db2SG

csdev95

root@csdev94: hares -list | more Infx csdev94 Infx csdev95 InfxDiskGroup csdev94 InfxDiskGroup csdev95 InfxS1_chk01_Volume csdev94 InfxS1_chk01_Volume csdev95 InfxS1_chk02_Volume csdev94 InfxS1_chk02_Volume csdev95 InfxS1_chk03_Volume csdev94 InfxS1_chk03_Volume csdev95 InfxS1_dump_Mount csdev94 InfxS1_dump_Mount csdev95 InfxS1_dump_Volume csdev94 InfxS1_dump_Volume csdev95 InfxS1_exports_Mount csdev94 InfxS1_exports_Mount csdev95 InfxS1_exports_Volume csdev94 InfxS1_exports_Volume csdev95 InfxS1_llog_Volume csdev94 InfxS1_llog_Volume csdev95 InfxS1_logdev_Mount csdev94 InfxS1_logdev_Mount csdev95 InfxS1_logdev_Volume csdev94 InfxS1_logdev_Volume csdev95 InfxS1_onunloads_Mount csdev94 InfxS1_onunloads_Mount csdev95 InfxS1_onunloads_Volume csdev94 InfxS1_onunloads_Volume csdev95 InfxS1_parameters_Mount csdev94 InfxS1_parameters_Mount csdev95 InfxS1_patameters_Volume csdev94 InfxS1_patameters_Volume csdev95 InfxS1_plog_Volume csdev94 InfxS1_plog_Volume csdev95 InfxS1_root_Volume csdev94 InfxS1_root_Volume csdev95 InfxS1_tapedev_Mount csdev94 InfxS1_tapedev_Mount csdev95 InfxS1_tapedev_Volume csdev94 InfxS1_tapedev_Volume csdev95 InfxS1_tmp01_Volume csdev94 InfxS1_tmp01_Volume csdev95 InfxS1_tmp02_Volume csdev94

InfxS1_tmp02_Volume csdev95Infx_IP Infx_IP csdev95 Infx_NIC csdev94 Infx_NIC csdev95 OraArchFS01_Mount csdev94 OraArchFS01_Mount csdev95 OraArchFS01_Volume csdev94 OraArchFS01_Volume csdev95 OraBkupFS01_Mount csdev94 OraBkupFS01_Mount csdev95 OraBkupFS01_Volume csdev94 OraBkupFS01_Volume csdev95 OraBkupFS02_Mount csdev94 OraBkupFS02_Mount csdev95 OraBkupFS02_Volume csdev94 OraBkupFS02_Volume csdev95 OraBkupFS03_Mount csdev94 OraBkupFS03_Mount csdev95 OraBkupFS03_Volume csdev94 OraBkupFS03_Volume csdev95 OraDataFS01_Mount csdev94 OraDataFS01_Mount csdev95 .

csdev94

root@csdev94: haclus -display #Attribute Value Administrators admin dm6757 km2847 AllowNativeCliUsers 0 ClusterLocation ClusterName AOTEST ClusterOwner CompareRSM 0 CounterInterval 5 DumpingMembership 0 EngineClass RT EnginePriority GlobalCounter 7281819 GroupLimit 200 HacliUserLevel NONE LockMemory ALL LogSize 33554432 MajorVersion 3 MinorVersion 5 Notifier State: none Host: none Severity: Information Queue: 30 Operators

PanicOnNoMem 0 PrintMsg 0 ProcessClass TS ProcessPriority ReadOnly 1 ResourceLimit 5000 SourceFile ./main.cf TypeLimit 100 UserNames admin dm6757 km2847 VCSMode VCS_7 6). Configuration Information: /etc/VRTSvcs/conf/config/main.cf To verify the contents of main.cf hacf verify /etc/VRTSvcs/conf/config 7). Config files:
root@csdev94: more /etc/llttab set-node csdev94 set-cluster 1 link eri0 /dev/eri:0 - ether - link qfe3 /dev/qfe:3 - ether - start root@csdev94: more /etc/gabtab /sbin/gabconfig -c -n2 root@csdev94: more /etc/llthosts 0 csdev94 1 csdev95

do not edit this file manually.

Things to remember: Applications are may not be setup as critical; dont want them to fail over automatically. All mounts need to be defined with fsckopt y To do any configuration changes, must 1st haconf makerw; make configuration changes..; haconf makero -dump

Other items: Note that before VCS can accept HA commands, the cluster nodes must be seeded. If the nodes are not seeded and you attempt to issue a command, you receive the error message: VCS:11037:Node has not received cluster membership yet, cannot process HA command Force feed a cluster: To start the cluster with fewer nodes than the total, verify the nodes to be excluded are actually down, then start GAB with the command /sbin/gabconfig -c -x. This manually seeds the cluster and enables VCS to start on all connected systems. Stale configuration: - Admin Wait state It is possible that VCS marked the configuration stale by creating a .stale file because the last VCS shutdown was not graceful. The .stale file is created in the directory /etc/VRTSvcs/conf/config. Type the following commands to verify the configuration and force VCS to reread the configuration file: # cd /etc/VRTSvcs/conf/config # hacf -verify . # hasys -force system Starting VCS using a stale flag 2 nodes in cluster, both in admin wait or stale state. # s1 has the good main.cf file you'll need to verify this hacf verify command # cluster software is not running on either node before executing the following: # hastart # hastart stale on s1 on s2

Dealing with faulted resources: hares -state looking for anything faulted - if faulted, clear with this

hares -clear {resource name} -sys {sysname} hares -clear pemsl01_fsweb -sys csppcon2 hagrp state looking for anything faulted

hagrp -clear GrpName -sys SysName

- if group faulted, clear with this

Вам также может понравиться