[jboss-jira] [JBoss JIRA] Updated: (JBMESSAGING-1706) JBoss Messaging sometimes causes a HeuristicMixedException

Wed Sep 23 02:19:49 EDT 2009

     [ https://jira.jboss.org/jira/browse/JBMESSAGING-1706?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Richard Kennard updated JBMESSAGING-1706:
-----------------------------------------

    Attachment: server.zip

My apologies for the delay in getting back to you. I am having a hard time creating a repeatable test case of this problem. However, in the meantime I have at least captured a log.

The log (attached) is:

1. With everything set to log level ALL. At least I hope it's 'everything' :)
2. From the state when the system is 'stuck' in the HeuristicMixedException described. The log does NOT show the first time the HeuristicMixedException occured on the cluster node (It is quite easy to trigger it the first time just by shutting down a node in the cluster, I do have a repeatable test case for this). Rather, the log shows a subsequent attempt to access a JMS queue and how it fails. All subsequent attempts on all queues on the node fail in the same way. I have to reboot the node to get out of this 'stuck' state
3. The problem starts around 2009-09-23 16:08:49,082

Please let me know if this offers any clues.

Regards,

Richard.

> JBoss Messaging sometimes causes a HeuristicMixedException
> ----------------------------------------------------------
>
>                 Key: JBMESSAGING-1706
>                 URL: https://jira.jboss.org/jira/browse/JBMESSAGING-1706
>             Project: JBoss Messaging
>          Issue Type: Bug
>          Components: JMS Clustering
>    Affects Versions: 1.4.4.GA
>            Reporter: Richard Kennard
>            Assignee: Howard Gao
>             Fix For: 1.4.0.SP3.CP10, 1.4.6.GA
>
>         Attachments: server.zip
>
>
> I am raising this JIRA to ask the JBoss Messaging Team to take ownership of this issue.
> The issue is that JBM intermittently (ie. when processing large amounts of messages, when bringing up/shutting down nodes in a cluster) throws a HeuristicMixedException. Once it does, that node in the cluster is effectively dead. No more messages get processed. My app has multiple queues, but once one queue on the node fails (say, the e-mail queue) then all queues stop working (say, the reports queues): they all report HeuristicMixedException whenever a messages is sent to them.
> I am using a workaround by having a separate monitoring program that reboots the node whenever it gets stuck in this HeuristicMixedException state.
> The JBoss ESB team have also had to implement a workaround: https://jira.jboss.org/jira/browse/JBESB-2484
> Others on the forum are seeing this too: http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4241134#4241134
> According to Kevin Conner of the ESB team, Howard worked with them on a workaround: https://jira.jboss.org/jira/browse/JBPAPP-1642, but the issue persists.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira