[
https://issues.apache.org/jira/browse/AMQ-3697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13201348#comment-13201348
]
William commented on AMQ-3697:
------------------------------
I am able to reproduce the problem. The scenario is the following.
Send a lot of messages, let's say hundreds very quickly on one topic with one
consumer. Messages are consumed and suddently (I think when the send rate is
too high) messages stopped to be consume. I run the debugger and I can see
that when the prefetch size is higher than the dispatched count the topic is
condiered as full and messages are not dispatched to the consumer.
The situation can only get back to normal when the consumer is restarted
because the counters change.
So Here are my questions :
* What does dispatchCount, dequeueCount, etc mean ?
* Do I have to adjust message acknowledgement and prefetch size ?
* Does the fact that dequeue count never increase means that the consumers are
blocked ?
> Messages not delivered to connecter consumer
> --------------------------------------------
>
> Key: AMQ-3697
> URL: https://issues.apache.org/jira/browse/AMQ-3697
> Project: ActiveMQ
> Issue Type: Bug
> Components: Broker
> Affects Versions: 5.5.1
> Environment: SunOS 5.10 Generic_142909-17 sun4v sparc SUNW,T5140
> Reporter: William
> Priority: Critical
>
> I am using ActiveMQ 5.5.1 and sometimes messages are not dispatched to
> connected consumers.
> A single broker is used. No transaction and no persistence is used for
> message delivery. Message consumers are spring template made for topic
> consuption. Optimize acknowledge is enabled and acknowledgment is auto.
> By Looking at JMX Topic MBean I can see that messages are received but not
> dispatched (and of course non depueued). I noticed that disptatched count was
> 100 more than dequeue count.
> From JMX :
> * Received count : 4502
> * Dispatched count : 3756
> * Dequeued count : 3656
> If a start a new consumer on the same topic, the messages are only dispatched
> to the new consumer while the older one is still seen as connected (consumer
> count on the mbean is 2). Once the problem appear the only solution is to
> restart the consumer or the activemq instance.
> It seems that even if the consumer is seen as connected, messages as not
> dispatched for an unknow reason.
> Which logs can I enable in order to diagnose the problem (on activemq side
> and/or consumer side) ? I have already made a thread dump, not thread are
> blocke
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
