Best practice to shutdown/startup when failing over. (4224559)

For cluster failover, shutting down does not stop cop completely, not read releasing shared memory segment and corrupted queues when it failed over to the other node.

Timing

May need modify the cluster script to use shutdown force or login as root to kill all shareplex processes during failover.

Shutting down:
1) shutdown listener
2) shutdown shareplex first since shareplex has user connections to oracle (use shutdown first, if not work then shutdown force, or kill as root)
3) shutdown oracle

Starting up:
Make sure shareplex var dir is mounted to the failover server, floating ip is pointing to the failed over server and oracle binaries etc are available on failed over node.
1) start database
2) start shareplex
3) start listener to allow user access