Hi, yes, Im working my way through the artemis it tests to model my own for testing.
What I do know now is that many of the mqtt messsages are ~22k size So what I did to reproduce: 1. artemis create <instance> 2. model any spec in artemis.profile (ex xmx/xms=1024M) 3. make a 22k test file 4. running a JMS consumer on <MyQueue> 5. for (( c=1;c<100000;c++)) do echo "Running $c....."; mosquitto_pub -t MyQueue -f ./data.txt; done 6. observe with visualvm. After a while (and exactly at the same count) artemis stops accepting packages, no more messages are sent and log statements: io.netty.channel.DefaultChannelPipeline] An exceptionCaught() event was fired, and it reached at the tail of the pipeline. It usually means the last handler in the pipeline did not handle the exception.: io.netty.util.internal.OutOfDirectMemoryError: failed to allocate 16777216 byte(s) of direct memory (used: 520595456, max: 536870912) Loops in an incredible speed. MessagesAcknowledged and MessagesAdded are the same for the testqueue. Not sure if it matters, but I can run, let's say 50% of the critical count, do a GC on artemis (via visuamvm) and start another test. I would expect ~100% then, but havoc's at ~50%. To me this should signal a memory leak. /hw -- View this message in context: http://activemq.2283324.n4.nabble.com/artemis-2-1-0-logging-disaster-tp4726922p4727028.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
