[Pacemaker] MySQL HA-Cluster, error on move

[Devin Reade]

I should also mention that some of the key pieces are having the
pid file on local (non-drbd) disk and having your db configured to
allow the RA to test the status (per the docs).  Having the pid
file on drbd storage causes the monitor on the backup node to fail.

I don't use a different db data location than the lsb script so
that if there is a node running an independent mysql instance
it doesn't affect the cluster through false positives in the 
monitoring.

Devin