I deployed two messaging nodes over separate systems. One on a Windows XP professional and another on CentOS (linux). I removed the bind_addr from postoffice config in clustered-mysql-persistence-service.xml. Then I ran the sample code for queue_faiover. Half times I got sucess: I received Helo1 first and Helo2 after failover. Other times I got both Helo1s before and after failover. But in all cases, I can see "sever side fail over is now complete" on messaging-node1.