[Pacemaker] Resources restarted when second node leave/join the cluster

Andrew Beekhof andrew at beekhof.net
Mon Jul 5 09:46:01 UTC 2010


On Fri, Jul 2, 2010 at 6:51 AM, levin <levin at mydream.com.hk> wrote:
> Hi,
>
> I have a two node cluster (A/S) running on a SuSE 11 box with clvm on SAN shared disk, it was found that a strange behavior on clone resource dependency(order) which will cause the whole resources tree restarted in the event of 1 cluster node leave/join the cluster.
>
> When a resource depends on a clone resource, for example "order order-vg01 inf: dlm-clvm-clone vg01", the "vg01" will restart in the event of status change on "dlm-clvm-clone" by 1 node leave/join.
>
> Anyone know how to prevent a restart by 1 node leave/join?

Update to SP1.  It should be fixed there.

>
> or
>
> In my case, would you suggest to put dlm/clvm into rcopenais(startup script on suse) instead of managed by cluster.
>
> Thanks alot
> Regards,
> Levin
>
>
> This packages I'm using --
>
> libopenais2-0.80.3-26.8.1
> openais-0.80.3-26.8.1
> libopenais-devel-0.80.3-26.8.1
> lvm2-clvm-2.02.39-18.11.17
> libpacemaker3-1.0.5-0.5.6
> pacemaker-pygui-1.99.2-0.2.6
> libpacemaker-devel-1.0.5-0.5.6
> pacemaker-1.0.5-0.5.6
> heartbeat-ldirectord-3.0.0-0.2.8
> heartbeat-common-3.0.0-0.6.5
> libheartbeat-devel-3.0.0-0.6.5
> heartbeat-resources-3.0.0-0.2.8
> libheartbeat2-3.0.0-0.6.5
>
>
> _______________________________________________
> 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
> Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
>




More information about the Pacemaker mailing list