I am implementing a cluster for Jboss 5.1 for my project specific configuration (instead of using 'all' configuration). I am currently doing a 2 node cluster on the same machine. My MDBs are deployed on both the nodes. We have queues which I have marked as clustered.
The problem is :
Even in normal processing (without a failure), the MDBs on both the nodes are picking up the job and going on to process it. This is leading to duplicate processing.
How do I ensure that only one of the MDBs picks up the message?