I'm having an issue with amq 5.1-SNAPSHOT, when stopping and restarting the broker.
I'll process a batch of messages through my application. I'll look on the webconsole, and see that all queues have no messages pending. I then shutdown the broker. Often, when I bring the broker back up, the webconsole shows that there are several messages back in the pending state. Sometimes, these get immediately redelivered to the application, sometimes they don't. Sometimes also, if I then submit more fresh messages, these get processed fine, but the ones that were pending on restart remain there. Finally, sometimes, the number of pending messages gets updated to a negative number. I'm using amqPersistence. But I saw the same thing happening too with journaled jdbc/derby persistence (I didn't test that scenario too thoroughly though).... Anyway, I'm wondering if there are any ideas about this, does anything obvious jump to mind. Are there any config params to try? I tried setting the syncOnWrite flag to true, that didn't help.... I'm wondering about the cleanupInterval param, might that be something to reduce (default is 3 mins)... Anyway, trying to understand this is a bug (seems like it) or the desired behavior with the default config....something to do with optimizing performance? The message redelivery is the main thing that I'm concerned about.... Jason -- View this message in context: http://www.nabble.com/problem-with-messages-%27pending%27-and-redelivery%2C-after-broker-restart-tp15363332s2354p15363332.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
