LVM resource agent becomes 'unmanaged' on one node

  • 3408390
  • 31-Aug-2006
  • 26-Apr-2012

Environment

Novell SUSE Linux Enterprise Server 10 (SLES10)
High Availability Release 2 (HA2)
Heartbeat
Linux Volume Management (LVM) Resource Agent

Situation

When an LVM resouce agent fails over to another node, the resource goes 'unmanaged' and fails to run on the node. The /var/log/messages shows the following errors:

Aug 25 16:18:27 han1 LVM[9599]: [9608]: INFO: Activating volume group san
Aug 25 16:18:27 han1 LVM[9599]: [9612]: INFO: File descriptor 3 left open File descriptor 4 left open Reading all physical volumes. This may take a while... Found volume group "san" using metadata type lvm2
Aug 25 16:18:28 han1 LVM[9599]: [9654]: INFO: File descriptor 3 left open File descriptor 4 left open 2 logical volume(s) in volume group "san" now active
Aug 25 16:18:28 han1 lrmd: [7425]: info: RA output: (webvg:start:stderr) File descriptor 3 left open
Aug 25 16:18:28 han1 lrmd: [7425]: info: RA output: (webvg:start:stderr) File descriptor 4 left open


Resolution

Turn boot.lvm off on all nodes, and reboot. The LVM Resource Agent will activate the volume on the correct node as needed. Run the following commands on all nodes to correct the problem:

chkconfig boot.lvm off
reboot

Additional Information

With boot.lvm turned on, the shared disk media are scanned at boot time for volume groups. All volume groups found are then activated. This creates a conflict with the LVM Resource Agent, when it attempts to activate the LVM volume and finds it's already active.