[ClusterLabs] How to set up fencing/stonith

[Casey & Gina]

> So, then instead of powering off the VM in vSphere, I instead tried a `killall -9 corosync` on the primary.  This resulted in the VIP coming up on node 3, and node 1 being rebooted.  Great!

Unfortunately, things don't work at all when it comes to the PostgreSQL resource agent...  When I `killall -9 corosync` on the primary node (node 3 in this case), the cluster ends up in the following state, with the VIP and PostgreSQL down indefinitely.  Node 3 is reset and comes back up, but as I don't have pacemaker and corosync set to autostart, nothing happens until I log on and issue a `pcs cluster start` on that node, which causes it to resume it's former primary status.  What do I need to do to make the resources fail over to one of the standby nodes when the primary goes offline?  I feel like I must be missing something obvious...

 vfencing       (stonith:external/vcenter):     Started d-gp2-dbpg0-1
 postgresql-master-vip  (ocf::heartbeat:IPaddr2):       Stopped
 Master/Slave Set: postgresql-ha [postgresql-10-main]
     Slaves: [ d-gp2-dbpg0-1 d-gp2-dbpg0-2 ]
     Stopped: [ d-gp2-dbpg0-3 ]

Thanks,
-- 
Casey