The way our application is set up right now is that: We have 200 consumers that can process jobs. A message is a single job (400k of data as noted above). On start up, the consumer connects to the advisory queue and monitors queue adds & deletes. When it gets a queue add notification, it starts consuming messages from that queue (via RecvNoWait()). It continues polling that queue until it gets a queue delete notification. The consumer does its work and sends a response back (also about 400k) via the corresponding response queue.
A Producer, upon start up creates 2 queues, a SendQueue and a Recv queue (the consumers ignore the recv queues obviously). It posts its jobs to the send queue and polls the response queue for responses. Upon a producer shutting down, it will delete the 2 queues it created. Right now, after about an hour or so, I have one producer that has sent/recv 43,000 msgs and one that has sent/recv about 1000. -- View this message in context: http://activemq.2283324.n4.nabble.com/Can-you-help-with-these-exceptions-tp4670752p4670773.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.