In a live-backup scenario, if the live is restarted and shutdown too soon,
the client have a chance to fail on failover because it's internal topology
is inconsistent with the final status. The client keeps connecting to live
already shut down, never trying to connect to the backup.

This fix should go with HORNETQ-1571.

Version-Release number of selected component (if applicable):
EAP 6.4 CP20

How reproducible:
Always

Steps to Reproduce:
1. Failover to backup server
2. Start live server
3. As soon as the live server starts processing messages, kill the live server

Actual results:

Clients will not fail back to the backup, even though failover is successful. They keep trying to connect to the live server which was killed.