[ 
https://issues.apache.org/activemq/browse/AMQ-2887?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=61738#action_61738
 ] 

C Velo edited comment on AMQ-2887 at 9/8/10 8:36 PM:
-----------------------------------------------------

I can also see the negative queue counts in JConsole as well.

Haven't been able to reproduce a test case with this using JMeter, however,
I have been able to reproduce this via using an in-house test script which
puts large #'s of our application-specific messages on the application queues,
which are then pulled off by application consumers which also do database work.

+Scenario:+
            • Producers: 3
            • Brokers (network of brokers): 2
            • Consumers (application consumers w/ connection pooling): 2

Our message sizes are small - all of them under 1k.

----

To reproduce, start up multiple Producers, both of which are throwing 80,000+ 
messages
at two brokers (which are networked together) and have the two consumers 
running on 
separate servers pulling items off of the application queue.

----

+NetworkConnector snippet:+
{code}
        <networkConnector name="ActiveMQ Cluster" 
uri="static:(tcp://broker1:61616,tcp://broker2:61616,tcp://broker3:61616)" \
conduitSubscriptions="false" \
dynamicOnly="true" \
decreaseNetworkConsumerPriority="false" \
suppressDuplicateQueueSubscriptions="true" \
duplex="true" \
networkTTL="5"/>
        </networkConnectors>
{code}



      was (Author: cvelo):
    I can also see the negative queue counts in JConsole as well.

Haven't been able to reproduce a test case with this using JMeter, however,
I have been able to reproduce this via using an in-house test script which
puts large #'s of our application-specific messages on the application queues,
which are then pulled off by application consumers which also do database work.

+Scenario:+
            • Producers: 3
            • Brokers (network of brokers): 2
            • Consumers (application consumers): 2

Our message sizes are small - all of them under 1k.

----

To reproduce, start up multiple Producers, both of which are throwing 80,000+ 
messages
at two brokers (which are networked together) and have the two consumers 
running on 
separate servers pulling items off of the application queue.

----

+NetworkConnector snippet:+
{code}
        <networkConnector name="ActiveMQ Cluster" 
uri="static:(tcp://broker1:61616,tcp://broker2:61616,tcp://broker3:61616)" \
conduitSubscriptions="false" \
dynamicOnly="true" \
decreaseNetworkConsumerPriority="false" \
suppressDuplicateQueueSubscriptions="true" \
duplex="true" \
networkTTL="5"/>
        </networkConnectors>
{code}


  
> Seeing negative numbers in the "Number Of Pending Messages" column in the web 
> console
> -------------------------------------------------------------------------------------
>
>                 Key: AMQ-2887
>                 URL: https://issues.apache.org/activemq/browse/AMQ-2887
>             Project: ActiveMQ
>          Issue Type: Bug
>    Affects Versions: 5.4.0
>         Environment: Linux 2.6.18-128.1.10.el5.xs5.5.0.51xen #1 SMP Wed Nov 
> 11 07:38:08 EST 2009 i686 i686 i386 GNU/Linux
>            Reporter: C Velo
>         Attachments: activemq_broker1.xml, pubd-amq3_negative_counts.pdf
>
>
> Seeing negative numbers in the "Number Of Pending Messages" column in the web 
> console:
> ibsys.content.publication     -9771   262     1192568 1202339

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to