Hi James, that's fair comment, apologies. My scenario is below :
Persistant Queues, using out of the box persistance. Message size approx 110kb. Max Message Volume on a requestQueue circa 50,000 1 Producer to a requestQueue 1 MDB (with say 10 threads configurable in Websphere AS 5.1) listening on requestQueue. MDBs pass message to Stateless Session Bean, which then retrieves data against a database, it is here that the delay of approx 9 seconds occurs. Based on the response from the database we can then deliver to an exceptionQueue, generalExceptionQueue or responseQueue. My application is J2EE1.3 spec and is deployed within WAS 5.1, as we can connect to any JMS queue provider our code in the Application is generic and we work off the QueueConnectionFactory to get to our queues. My main problem is that I need a configuration of ActiveMQ whereby I can set the prefetch limit to something really small, so if I have 10 MDBS perhaps only a prefetch limit of 10 and I guess my frustration lies here, as I'm not coding the connection to the queue specifically. Should I set the consumer.preFetchsize=1 when configuring the queues in Websphere ? Or should I use the jms.preFetchPolicy.queuePrefetch=1 setting ? Can I define my queues in ActiveMQ.xml and give them the above settings ? As you have gleaned already I need a configuration that will simply Accept messages onto a queue, which will build up to a volume of circa 50000 messages, and then have my consumer process them as and when it can, without losing messages. Again, apologies for my rant. Muzza. -- View this message in context: http://www.nabble.com/Slow-consumer-configuration-tf2106258.html#a5810362 Sent from the ActiveMQ - User forum at Nabble.com.
