[
https://issues.apache.org/activemq/browse/AMQ-1187?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_39545
]
Tom Samplonius commented on AMQ-1187:
-------------------------------------
The negative queue size bug is 100% reproducible, and affects JMX and the Web
Console counters. Just put a 100 persistent messages into a queue, stop
ActiveMQ, and then start ActiveMQ, and tell a consumer to read all of the
messages. The queue size will be "-100". ActiveMQ does not ask its
persistence engine to return the number of messages already in the queue, so it
always starts with all counters at zero.
Now, JConsole may lag. I don't know how often it refreshes. But it doesn't
not sound like you are having the negative queue size problem, unless the OP
started and stopped ActiveMQ with messages in the queue.
> Messages sent in parallel threads may not be enqueued
> -----------------------------------------------------
>
> Key: AMQ-1187
> URL: https://issues.apache.org/activemq/browse/AMQ-1187
> Project: ActiveMQ
> Issue Type: Bug
> Affects Versions: 4.0.2, 4.1.0
> Environment: Linux
> Reporter: Philippe Göllnitz
> Attachments: console.txt, jconsole.png, TestMqBroker.java
>
>
> I sent 1000 messages in own threads but monitoring the queue with jconsole
> sometimes only 999 are in the queue.
> This happens more likely if you raise the number of sent messages 10.000 or
> 100.000 but can also happen if set to 100.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
