Error Message

2010/09/13 14:10:57 VCS WARNING V-16-10011-10125 (node2) LVMVG:lvmvg_vg2:online:The child process (/usr/sbin/lqueryvg -p hdisk22 -T ) failed to get the VGDA timestamp

2010/09/13 14:10:57 VCS WARNING V-16-10011-10107 (node2) LVMVG:lvmvg_vg2:online:Failed to varyoff the volume group testvg for sync. Sync failed. Errno is Error 0

2010/09/13 14:12:58 VCS ERROR V-16-2-13066 (node2) Agent is calling clean for resource(lvmvg_vg2) because the resource is not up even after online completed.

If physical volumes are added, deleted or replaced in a volume group, the disk entries in the  ODM database of the local nodes are updated, but some or all of the ODM entries on other nodes remain stale. This causes the volume group resource online failure on those nodes.

This is tracked via Symantec internal incident # 2190134.

Proposed solution entails introduction of a new temporary attribute as well as an action entry point to LVMVG agent. From a user's perspective, the following steps are to be followed on the node where the resource is active:

1. migratepv (to migrate to new disks)
2. reducevg (to delete old disks)
3. Invoke the newly introduced VCS action entry point

A new action entry point updatepv is introduced in the LVMVG agent to resolve this issue. The updated disk information having the list of physical volume IDs stored in a temporary file can be propagated to all nodes in the cluster using the updatepv action entry point. The agent uses this disk information to bring the volume group resource online during a failover. Once the volume group is online on the failover node, the ODM entries get updated and then the temporary file is deleted.

This change has been included in VRTSvcsag 5.1SP1RPHF2 on AIX platform.

Please contact Veritas Enterprise Support to obtain this hotfix.

Applies To

Systems running:

- AIX Operating System

- VCS cluster running any patch levels in 5.0 or 5.1

- VCS resources configured with LVMVG resources