Re: Stress testing: OOME

2014-02-14 Thread jlindwall
I'll be re-running my stress test and Monday and report my findings back to this thread. Thanks! Your questions are thought provoking. -- View this message in context: http://activemq.2283324.n4.nabble.com/Stress-testing-OOME-tp4677959p4678010.html Sent from the ActiveMQ - User mailing list ar

Re: Stress testing: OOME

2014-02-14 Thread jlindwall
Thanks all for the comments - I will rerun my stress test on Monday and report back. I do not think my url-based parameter were being used because of pilot error: I was using a jndi.properties file (a practice I have now abandoned) and the connection url was specified in that file, not in my exter

Re: Stress testing: OOME

2014-02-14 Thread Christian Posta
Want to post the code for your consumer? Is each consumer in its own JVM or are all consumers sharing the same JVM? Sharing same Session/Connection? On Thu, Feb 13, 2014 at 2:31 PM, jlindwall wrote: > I am stress testing activemq 5.9.0 by sending a flood of non-persistent > messages to a topic us

Re: Stress testing: OOME

2014-02-13 Thread artnaseef
Wait a second. Is that 100 consumers all running in the same JVM? Keep in mind that consumers of a Topic all get a copy of every message - they are not load-balancing. So, if the producers are sending 1000 msg/sec and there are 100 consumers, that 10,000 msg/sec attempting to be consumed. Still

Re: Stress testing: OOME

2014-02-13 Thread artnaseef
First thing is finding that memory - if visualvm doesn't show much on the HEAP, how about permgen? Also, how about the CPU activity graph for GC? Does it spike? Based on the description so far, there's no valid reason for the conusmer JVM to be OOM unless its heap is just extremely small. --

Re: Stress testing: OOME

2014-02-13 Thread jlindwall
Thanks for the tip. Is this documentation still accurate? http://activemq.apache.org/what-is-the-prefetch-limit-for.html I tried setting the limit on the connection url but got the same behavior: tcp://testapp01:61616?jms.prefetchPolicy.all=50 I also tried setting the limit on my dest

Re: Stress testing: OOME

2014-02-13 Thread Timothy Bish
On 02/13/2014 04:31 PM, jlindwall wrote: I am stress testing activemq 5.9.0 by sending a flood of non-persistent messages to a topic using jmeter. My producers are 50 threads attempting to deliver 1000 msgs/sec; each of size 1K. There is merely a 1 millisecond delay in between message sends. Th