[
https://issues.apache.org/activemq/browse/AMQ-2962?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=62529#action_62529
]
Gary Tully commented on AMQ-2962:
---------------------------------
The stack trace, (in the hung thread dump) is odd, seems to be stuck on a
reentrant writelock lock call, when it already has the writeLock, so it is as
if the reentrant bit is not working!
Can you try the sun jdk?
> Hanging at startup
> ------------------
>
> Key: AMQ-2962
> URL: https://issues.apache.org/activemq/browse/AMQ-2962
> Project: ActiveMQ
> Issue Type: Bug
> Components: Message Store
> Affects Versions: 5.4.1
> Environment: Fedora Core 12
> Reporter: Roelof Naude
> Priority: Minor
> Attachments: debug_traces_hanging.log, debug_traces_startup.log,
> thread_dump_during_hang.txt
>
>
> Noticed today the activemq was hanging during startup. Investigation revealed
> that it is tied to expiration of messages at startup. The only way to recover
> from this was to kill -9 the VM and delete the kahadb message store.
> Further testing was done whereby message expiry time was set to longer
> period. AMQ started up just fine. Another test was performed whereby:
> 1. subscribe for messages
> 2. send a few messages
> 3. disconnect consumer
> 4. send more messages (1 or 2 will do)
> 5. shutdown AMQ
> 6. wait for a time period exceeded the expiration time of the messages
> 7. start AMQ. at this time it should be hanging
> Logs and thread dumps attached.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
