On Thu, May 28, 2009 at 4:28 PM, firewall <gab...@gmail.com> wrote: > > Hi all, > > I'm having problems getting a split-brain to work (for demo, yes that's > right). Basically, I have an active/passive cluster both configured with > pingd like below. When the ping node is not reachable on the primary/active > (i.e. I unplug the cable..), a failover occurs to the backup and vice-versa. > However, I expect that when the ping node is not reachable from both nodes, > a split-brain should occur resulting in the cluster resource running on both > nodes at the same time. I need this to happen for demonstration.
You actually want this to occur?!? First of all, split-brain occurs when the two nodes can't see each other, not when they can't see the ping node. Second, the design goal for cluster software is to _prevent_ both sides running the same resources at the same time at all costs. I really question the use-case here. > I'm pretty > sure it's to do my config but not sure what to change so that a split-brain > would occur when both cluster nodes cannot ping the ping node. My cib.xml > file is attached and my ha.cf file is displayed below. > > http://www.nabble.com/file/p23762887/cib.xml cib.xml > gabbep > [snip] > #auto_failback yes > > PS. surprisingly, a failback still occurs albeit it's commented out in the > ha.cf file. That setting has no meaning in a crm cluster. You need to look up resource-stickiness in the documentation. _______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems