Please define concretely "Acking happens as needed." I strongly suspect based on what you've written that you're wrong and acking is in fact not happening as it needs to.
Do you understand how round-robin dispatch works for a queue with your two consumers? Do you understand what a consumer's prefetch buffer is and why you won't see messages on the web console if they're dispatched but unconsumed? Do you understand what happens to messages in a consumer's prefetch buffer if you don't ack a message in CLIENT_ACK mode? Based on what you've described, I think your entire paradigm is broken and you should have two separate queues, with your translator consuming on A and publishing to B. And I disagree with the first line of your original post: you do not have a simple setup. It's not the world's most complicated, but it's not simple. Tim On Dec 14, 2015 12:13 PM, "dabayev" <daba...@fortress.com> wrote: > Acking happens as needed. > > my page size is set to <policyEntry queue=">" maxPageSize="5000" > maxBrowsePageSize="5000"/> > > I have 923 messages on the queue. Any other ideas? > > > > -- > View this message in context: > http://activemq.2283324.n4.nabble.com/Issue-with-MessageConsumer-and-MessageSelector-tp4704870p4704939.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. >
