"running into a deadlock etc."
First, let me say ACTIVEMQ 5.1.0 Is GREAT!
Now for some free information:
I have been working on a RT/SOA w/ActiveMQ for the past few months, under
very heavy performance load, w/many queues and many competing
consumers,producers, and such; AMQ5.0&&AMQ5.1 will deadlock.
After much pain, and re-configuring; I can say for ceratin WHAT FIXED THE
PROBLEM FOR ME :)'. I noticed
the Journals would never get past 3-6 iterations of 32MB. I started
planning with the pendingQueuePolicy
and noticed once I switched the queues to a fileQueueCursor, everything
works perfect now! I have
presented an example below. I would think a simple junit will produce
similar results; sorry, I would like to make one; moreover deadlines and
such. Anyway; Great work, and I really enjoy your
product the more I work w/it.
Best regards
JRadecki
<policyEntry
queue="my.queue.name.none.of.your.businessConsumerFactory"
producerFlowControl="false" memoryLimit="10mb">
<dispatchPolicy>
<strictOrderDispatchPolicy/>
</dispatchPolicy>
<subscriptionRecoveryPolicy>
<lastImageSubscriptionRecoveryPolicy/>
</subscriptionRecoveryPolicy>
<deadLetterStrategy>
<individualDeadLetterStrategy queuePrefix="DLQ."/>
</deadLetterStrategy>
<pendingQueuePolicy>
<fileQueueCursor />
</pendingQueuePolicy>
</policyEntry>
--
View this message in context:
http://www.nabble.com/-jira--Created%3A-%28AMQ-1490%29-Deadlocks-%28with-JUnit-tests%29-tp13702734s2354p14605939.html
Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
