[Pacemaker] Pacemaker failover problem

[Andrew Beekhof]

On Tue, Mar 9, 2010 at 4:04 PM, Erich Weiler <weiler at soe.ucsc.edu> wrote:
> Thanks for the reply!  Yes, I have checked that my LSB scripts are
> compliant.

I'm pretty sure it fails step 3 of
   http://www.clusterlabs.org/doc/en-US/Pacemaker/1.0/html/Pacemaker_Explained/ap-lsb.html

> If this can provide any insight, here is a clip from
> /var/log/messsages on genome-ldap2 when genome-ldap1 goes down (when the
> order constraint is in place):

Please send logs as attachments.

> I also thought the expected behavior was to stop LDAP first, start the IP,
> then restart LDAP.  I think I'd rather have it behave that way.

Its possible you were affected by a clone ordering bug that was fixed in 1.0.8
But still, double check your LSB script... starting something that is
already running should not result in an error.