[Pacemaker] Unable to bring up cluster after crash and update

[Jonathan deBoer]

Hello all:

I am trying to get a corosync-openais-pacemaker cluster stack installed
and running on a 2 node cluster.

I am currently using the latest corosync and openais from the svn
repository (as of today, July 22)
The version of pacemaker is the 1.0 tip from July 19th.

Corosync has been crashing randomly on one of my nodes (The same one,
consistently). Today, after a crash, I shut down the cluster to upgrade
corosync hoping to resolve the problem.  Now pacemaker is causing a
crash on startup, and I cannot start either node.

I have attached the debug output from one node in the cluster. Is there
something I can do to reset the cluster state so it will start? Or is
this a bug?

Now that Pacemaker will build against openais 1.0.0 and corosync 1.0.0,
should I downgrade to those versions instead of using the latest in SVN?

What do you recommend (version wise) as a stable stack?

Any help resolving this would be greatly appreciated.

Thanks!

Jonathan deBoer
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: crash.log
URL: <http://lists.clusterlabs.org/pipermail/pacemaker/attachments/20090722/02f97d1b/attachment.log>