[Pacemaker] sles11, ocfs2, 2 node cluster with one storage, failover problem

[Dejan Muhamedagic]

Hi,

On Mon, May 17, 2010 at 12:24:12PM +0200, Patrik.Rapposch at knapp.com wrote:
> 
> hy,
> 
> I have following problem:
> 
> I configured a 2 node cluster running SLES11 with the HAE extension. I use
> "pacemaker-1.0.3-4.1", "openais-0.80.3-26.1" and "ocfs2 1.4.1".
> I used the SLES high availabilty guide to configure my cluster.
> 
> I have 2 ocfs2 mount clone-resources running on the cluster.
> 
> My first problem was, that the cluster didn't work in any way, because I
> defined 2 rings in the openais.conf, which only will be supported from
> novell with the SLES SP1, so I switched back to one ring.
> 
> The major problem I have now is following:
> 
> When I reboot one node, or stop openais on one node, the cluster behaves as
> it should, that means the surviving node
> is still able to access the ocfs2 mounts and works normaly.
> 
> But if I trigger an ifdown command on the interface, which is defined in

openais/pacemaker can't handle ifdown. Neither can corosync with
multiple rings. You have to make sure that the interface stays up
at all times (i.e. don't use dhcp).

Thanks,

Dejan

> the openais.conf, the node hangs and restarts after a while. The surviving
> node doesn't support to open any other ssh sessions and if I try to open a
> mounted directory, the existing session hangs and I have to wait, till the
> first node is up again.
> 
> I don't have a glue how this can happen and hope, that you can help me.
> 
> Attached you can find my configs and failure logs.
> 
> (See attached file: config_plus_failure_logs.rtf)
> 
> Mit freundlichen Grüßen / Best Regards
> 
> Patrik Rapposch, Bsc.
> Systemadministration
> 
> KNAPP Systemintegration GmbH
> Waltenbachstraße 9
> 8700 Leoben, Austria
> Phone: +43 3842 805
> Mobil:
> Fax: +43 3842 82930-990
> patrik.rapposch at knapp.com
> www.KNAPP.com
> 
> Commercial register number: FN 138870x
> Commercial register court: Leoben


> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf