Вы находитесь на странице: 1из 3

1/9/2021 Document 2189857.

1
Copyright (c) 2021, Oracle. All rights reserved. Oracle Confidential.

Solaris Cluster Failover LDom Resource Creation Fails with "Master-slave dependency
between primary domain and guest domain does not exist." (Doc ID 2189857.1)

APPLIES TO:

Solaris Cluster - Version 3.3 and later


Oracle Solaris on SPARC (64-bit)
Checked for relevance on 12-03-2018

SYMPTOMS

Cannot create failover LDom resource in Solaris Cluster environment.

Commands run:

# clrg create guest-ldom10-rg


# clrg online -M -n control-ldom1 guest-ldom10-rg

# clrs create -g guest-ldom10-rg -t SUNW.HAStoragePlus:10 -p


GlobalDevicePaths=/dev/global/dsk/d10s0 guest-ldom10-has

# clpstring create -b guest-ldom10-ldom -t resource guest-ldom10-rg_guest-ldom10-ldom_ldompasswd


Enter string value:
Enter string value again:

All looks ok then run the following command which fails:

# clresource create -g guest-ldom10-rg -t SUNW.ldom:10 -p Domain_name=guest-ldom10 -p


resource_dependencies=guest-ldom10-has guest-ldom10-ldom
clresource: (C189917) VALIDATE on resource guest-ldom10-ldom, resource group guest-ldom10-rg,
exited with non-zero exit status.
clresource: (C720144) Validation of resource guest-ldom10-ldom in resource group guest-ldom10-rg
on node control-ldom1 failed.
clresource: (C891200) Failed to create resource "guest-ldom10-ldom".

The /var/adm/messages show:

control-ldom1 Cluster.RGM.global.rgmd: [ID 515159 daemon.notice] method


<hastorageplus_prenet_start> completed successfully for resource <guest-ldom10-has>, resource
group <guest-ldom10-rg>, node <control-ldom1>, time used: 0% of timeout <1800 seconds>
control-ldom1 Cluster.RGM.global.rgmd: [ID 224900 daemon.notice] launching method
<hastorageplus_monitor_start> for resource <guest-ldom10-has>, resource group <guest-ldom10-rg>,
node <control-ldom1>, timeout <90> seconds
control-ldom1 Cluster.RGM.global.rgmd: [ID 515159 daemon.notice] method
<hastorageplus_monitor_start> completed successfully for resource <guest-ldom10-has>, resource
group <guest-ldom10-rg>, node <control-ldom1>, time used: 0% of timeout <90 seconds>
control-ldom1 Cluster.RGM.global.rgmd: [ID 224900 daemon.notice] launching method <gds_validate>
for resource <guest-ldom10-ldom>, resource group <guest-ldom10-rg>, node <control-ldom1>, timeout
<300> seconds
control-ldom1 SC[SUNW.ldom: 10,guest-ldom10-rg,guest-ldom10-ldom]: [ID 267012 daemon.error]
Master-slave dependency between primary domain and guest domain[guest-ldom10] does not exist.
control-ldom1 SC[,SUNW.ldom: 10,guest-ldom10-rg,guest-ldom10-ldom,gds_validate]: [ID 254920
daemon.error] Validate_command has failed rc=1
control-ldom1 Cluster.RGM.global.rgmd: [ID 699104 daemon.error] VALIDATE failed on resource
<guest-ldom10-ldom>, resource group <guest-ldom10-rg>, time used: 3% of timeout <300, seconds>

CHANGES
https://support.oracle.com/epmos/faces/DocumentDisplay?_adf.ctrl-state=qnrpgf94q_287&id=2189857.1 1/3
1/9/2021 Document 2189857.1

CAUSE

Failure policy settings on the primary domain was not set correctly.
The failure policy settings on the primary domain do not have the required failure policy setting. It must be set to 'failure-
policy=reset' on all the primary domains

This is documented in

Oracle Solaris Cluster 4.3

» Oracle Solaris Cluster Data Service for Oracle VM Server for SPARC Guide

» Installing and Configuring HA for Oracle VM Server

» SPARC: Planning the HA for Oracle VM Server Installation and Configuration

» SPARC: How to Configure Oracle VM Server for SPARC to Reset for Control Domain Failures

SOLUTION

Example from documentation referenced above.

Perform the following steps on the node where the ldg0 logical domain is created:

1. Set the domain failure policy to reset.

# ldm set-domain failure-policy=reset primary


# ldm list -o domain primary
NAME STATE FLAGS UTIL
primary active -n-cv- 0.6%

SOFTSTATE
Solaris running

HOSTID
0xXXXXXXXX

CONTROL
failure-policy=reset

DEPENDENCY
master=

2. Configure the master-slave domain dependency.

# ldm set-domain master=primary ldg0

3. Set the auto-boot? variable to false.

# ldm set-var auto-boot?=false ldg0


# ldm list -o domain ldg0
NAME STATE FLAGS UTIL
ldg0 active n--- 0.1%

SOFTSTATE
Solaris running

HOSTID

https://support.oracle.com/epmos/faces/DocumentDisplay?_adf.ctrl-state=qnrpgf94q_287&id=2189857.1 2/3
1/9/2021 Document 2189857.1
0xXXXXXXXX

CONTROL
failure-policy=reset

DEPENDENCY
master=primary
VARIABLES
auto-boot?=false
boot-device=vdisk1
keyboard-layout=US-English

Didn't find what you are looking for?

https://support.oracle.com/epmos/faces/DocumentDisplay?_adf.ctrl-state=qnrpgf94q_287&id=2189857.1 3/3

Вам также может понравиться