I was hoping someone might have an idea for how to handle duplicate detection when working with redistribution. Currently, it's throwing out messages.
Here's what we have:
-HornetQ 2.1.2 Final
-2 HornetQ Nodes, node A and B
-HornetQ is configured using a TCP cluster using statically-assigned nodes.
-HornetQ is configured to redistribute messages between A and B when one has consumers.
-The client is accessing these nodes via HA-JNDI (server1:port,server2:port).
Here's the issue:
1) I put 1000 messages on Node A
2) I restart Node A's jboss (at this point, all consumers connect to Node B via HA-JNDI)
3) All messages are redistributed to Node B
4) I restart Node B's jboss (at this point, all consumers connect to Node A via HA-JNDI)
5) All messages are redistributed to Node A
6) Node A recognizes these messages as duplicates and discards.
If I do NOT enable duplication detection, then I will occasionally get 1 or 2 messages delivered twice. If I DO enable duplication detection, then the 'fail back' throws away messages.
It's not that I think this is a bug, but I would like to know if there's something I should try differently.
You should probably play with Failover / backup nodes once 2.2 is out.