JMS Grid configuration

How is JMS grid configuration done for active-active failover?

Here's the test scenario. I have a simple jcaps project which has a jms topic and queue called jmsTopic and jmsQueue. The project is deployed (duplicated) to two different machines.

Each machine has a JMS grid daemon, within a cluster.

When I kill one of the machines in the middle of processing a set of messages I am expecting the other machine to take over the processing of messages. However, this is not happening. (If the machine is brought back up processing continues.)

Is there some configuration I am missing or are my expectations incorrect?

Regards,

Rupert

[677 byte] By [moonsita] at [2007-11-27 4:25:06]