[Pacemaker] OCFS & Pacemaker hang

[Lars Marowsky-Bree]

On 2009-07-17T14:21:22, Karl Katzke <kkatzke at tamu.edu> wrote:

> Jul 17 12:37:02 app-03 openais[5009]: [crm  ] ERROR: route_ais_message: Child 7098 spawned to record non-fatal assertion failure line 1299: dest > 0 && dest < SIZEOF(pcmk_children)
> Jul 17 12:37:02 app-03 openais[5009]: [crm  ] ERROR: route_ais_message: Invalid destination: 0
> Jul 17 12:37:02 app-03 openais[5009]: [MAIN ] Msg[24] (dest=local:unknown, from=app-05:crmd.5078, remote=true, size=901): <create_request_adv origin="send_direct_ack" t="crmd" version="3.0.1" subt="request" refer
> 
> That message was on the troublesome node. 
> 
> There was a discussion addressing this back in May, but I didn't see any resolution: http://lists.linux-ha.org/pipermail/linux-ha/2009-May/037254.html
> 
> Removing all of the ocfs-related resources and disabling ocfs2 on boot resolves the issue. 
> 
> Any suggestions? This is with the current SLES11-HAE versions of pacemaker, ocfs2, etc. I'm managing dlm and o2cb's startup and the mounting of all filesystems via the crm so that dependencies are met. 

If you're running SLE HA 11, the best suggestion I can make is to file a
bug report with Novell, providing hb_report from your cluster.

If you're not running SLE HA 11 or feeling adventurous, you could try
updating to the server:ha-clustering:UNSTABLE packages from the Build
Service to take a peek at the current development versions.


Regards,
    Lars

-- 
Architect Storage/HA, OPS Engineering, Novell, Inc.
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde