[jboss-jira] [JBoss JIRA] Reopened: (JBAS-8124) Memory Leak in JBossMQ with 2PC QueueSessions

Jason Greene (JIRA) jira-events at lists.jboss.org
Thu Apr 21 14:26:36 EDT 2011


     [ https://issues.jboss.org/browse/JBAS-8124?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jason Greene reopened JBAS-8124:
--------------------------------



Due to feedback from the community, I have split the AS7 and AS6 projects and reopened all unscheduled AS6 issues that are a year or less old. This will make it easier community members to find and work on them.

Future releases beyond 6.1 can be done provided a community member steps up to coordinate them.

> Memory Leak in JBossMQ with 2PC QueueSessions
> ---------------------------------------------
>
>                 Key: JBAS-8124
>                 URL: https://issues.jboss.org/browse/JBAS-8124
>             Project: Legacy JBoss Application Server 6 
>          Issue Type: Bug
>      Security Level: Public(Everyone can see) 
>          Components: JMS (JBossMQ)
>    Affects Versions: JBossAS-4.2.3.GA
>            Reporter: Rico Neubauer
>            Assignee: Adrian Brock
>             Fix For: No Release
>
>         Attachments: JBAS-8124.diff, MAT.png
>
>
> JBossMQ has a memory leak, which has the same symptoms as reported for JBoss Messaging in https://jira.jboss.org/browse/JBMESSAGING-638 and discussed in http://community.jboss.org/thread/128250.
> Problem is that when using a transacted QueueSession, when closing QueueSession#close invokes QueueSession#internalRollback to clean-up any unacknowledged messages, but also starts a new transaction, which is added to QueueConnection's SpyXAResourceProvider.
> Since this transaction is never committed or rolled-back, it stays for the lifetime of the QueueConnection, which can be like forever when cacheing the QueueConnection. Also see "Steps to Reproduce"
> Will attach MAT screenshot.
> This occurs in JBossMQ of 4.2.3, but also 4.0.5 (and probably all 4.x)
> I am aware of the EOL of JBoss 4.x, but will attach a proposed patch anyways.
> Possible workarounds:
> - Do not use transacted QueueSessions
> - Close and nullify the QueueConnection from time to time to let GC clean-up the leaked objects

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira


More information about the jboss-jira mailing list