MC/ServiceGuard

[Danny Howard]

Why does the message "cmclconfd: Failed to release volume group xxxxxxx" appear when I make changes to the SG cluster or packages. Message also appears when make_recovery is run on systems in a SG cluster. Any ideas?

[Rita C Workman]

It could be a couple things...but the first thing I would check is that /var/adm/cmcluster/.cm_start_file exists.

Check this thread:
http://us-support3.external.hp.com/cki/bin/doc.pl/sid=880ffd9f1820ddf5f2/screen=ckiDisplayDocument?docId=200000029146455

/rcw

[Danny Howard]

The file .cm_start_time is present on all systems in the cluster.

[melvyn burnard]

This is an "extraneous" message and can safely be ignored.
I believe it could be fixed by ensuring you have the latest LVM and SG patches, but am not 100% certain:-}

[Carsten Krege]

I couldn't read the document Rita referred to, but I disagree that this has anything to do with the .cm_start_file. :) (perhaps someone can quote the document?)

During cmapplyconf/cmcheckconf the cmclconfd opens all LVM physical volumes using LVM system calls. cmclconfd tries to query the cluster flags of the VG to see whether the VG is cluster aware. When cmclconfd is done it deattches the physical volumes again. If this deattach fails, the message is logged.

This message can usually be ignored.

Exception: If the message "failed to release volume group" is logged AND the message "WARNING: Cluster lock on disk /dev/dsk/cYtYdZ is missing!" comes up after start of the cluster (cmruncl). This is an open problem which HP is currently working on and where an workaround exists.

[Rita C Workman]

Here's the document: Doc ID = A5443251

Problem Description

My ServiceGuard cluster fails to start. I receive this message:

cmclconfd: Unable to lookup any node information in CDB: No such file or directory.
What is the problem?

Configuration Info
Operating System - HP-UX Version - 10.20
Hardware System - HP 9000
Series - K210

Solution

This problem has been known to occur when a cluster
node is missing /var/adm/cmcluster/.cm_start_time.

Solve the problem by recreating this file on the node that no longer has a copy.
**************************

/rcw