JMS failover

[gunjan_iitk]

Hi All,
Can anyone please advice me on how JMS works in the cluster setup.
How queues are maintained like do we have queues on each node, if so when one quehe fails then can MDBs on this ndoe can acess queue on other node.
Please help me I am new to jboss clustering.
Thanks and Regards,
Gunjan

[jason1]

destinations are on a single, MasterNode. When that fails, the HASingleton service redeploys them on another node

[mchizmar]

I have been trying to figure out how JMS and MDBs are handled during fail over as well and am confused by this answer. I understand the HASinlgeton process and this answer would make sense, But viewing other posts imply that MDB's are not supported by JBoss clustering, like the following :

http://www.jboss.com/index.html?module=bb&op=viewtopic&t=54596

Then I checked the JMS posts as well and was not able to find answers on how it works within a clustered env.

So, my questions are,
1. Are MDB's supported within clustering?
2. If so, then does anyone know of some documentation on how to configure the message beans and it's queues in order to get them to work in the clustered environment?

I am using 3.2.5 and using xdoclet annotations (if that is of importance) with a cluster of two nodes within the default partition.
Currently all non-message beans have been clustered fine; however, message beans and their queues are local to both nodes therefor two separate copies of each.

Thank you