After resolving my previous problem with the a network connector reconnection after a network outage I am now seeing duplicate messages being delivered from a queue. The situation is this:
I have a spoke instance which has a network connector to a hub instance. For testing purposes I have them running on different ports and am using a port forwarder to allow me to simulate a network outage. I have a queue producer connect to the spoke instance and as fast as it can send messages. Those are routed through the network connector to the hub instance and consumed there. While messages are flowing, the port forwarder is killed and brought back simulating an outage. If I send 1000 messages I always see 1000 messages being received by the hub's consumer. However when I look at the hub queue's JMX bean it says that there are a few extra messages still in the queue that have not been delivered to the consumer. These are duplicates that are apparently stuck in some state in the hub's broker. That have apparently been correctly identified as duplicates because they are not delivered, but neither have they been discarded. I was about convinced that it was just a problem with the JMX bean's statistics until I restarted the hub instance and reconnected the consumer. The duplicate messages were then delivered. Are there any settings that I can use in 5.5.1 to allow these duplicate messages to be discarded? I see some discussion of features in 5.6 that may help. I tried back-porting the auditNetworkProducers setting but either I didn't get all of the code changes or the setting isn't applicable in this case. Thanks, Bruce -- View this message in context: http://activemq.2283324.n4.nabble.com/AMQ-5-5-1-Duplicate-Messages-After-Network-Outage-and-Broker-Restart-tp4520697p4520697.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
