Could anyone confirm my findings or is this a bug/misuse I am experiencing.
Assume you have a consumer that handles messages from a producer. This consumer also sends heartbeats to some surveillance service at regular intervals. A producer sends messages to this consumer in such a pace that the amq broker queue gets full and producer throttling starts, usagemanager.waitforspace, 100% etc. Now if a heartbeat is sent while the producer is temporarily blocked everything stops!? The consumer handles the already received 1000 messages (prefetch) but is not given any more since the broker thinks that messages are not acknowledged, though it seems so.. Why is the message dispatching stopped in this case / messages not being acknowledged? The consumer has one connection, two threads and two sessions. Regards /Stefan ps. more info: http://www.nabble.com/Producer-and-consumer-hangs-with-specific-scenario-tf4492569s2354.html ds. -- View this message in context: http://www.nabble.com/Conections%2C-sessions-and-threads-tf4513656s2354.html#a12873896 Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
