One of my clusters (fairly old; heartbeat 3.0.0 on openais) runs VMs on Xen, with storage being directly on the drbd device. I noticed a similar problem where manual failovers were fine but cluster-initiated failovers failed. It turned out that setting a start delay on the xen RA solved the problem:
op start interval="0" timeout="60" start-delay="45" I never did any experimentation to determine the minimal working start-delay value, since it was intermittent to begin with and I just wanted it to work. That does mean, however, that it takes a while to fail over the resource. Not only do you have to wait for the VM OS to boot and its services to start, but there is also the delay before actually starting the VM. Devin -- If books were designed by Microsoft, the Anarchist's Cookbook would explode when you read it. - Mark W. Schumann _______________________________________________ Pacemaker mailing list: Pacemaker@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
