[infinispan-issues] [JBoss JIRA] Commented: (ISPN-1127) Memory leak in CacheStoreInterceptor
Sergey Kondratyev (JIRA)
jira-events at lists.jboss.org
Wed May 25 23:33:01 EDT 2011
[ https://issues.jboss.org/browse/ISPN-1127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12604493#comment-12604493 ]
Sergey Kondratyev commented on ISPN-1127:
-----------------------------------------
Try to start the test and when test sleeps do
jmap -histo:live PID-HERE|grep org.infinispan.transaction.xa.GlobalTransaction
It shows 10k transaction objects for me (the same in memory dump if created).
Hope this helps.
> Memory leak in CacheStoreInterceptor
> ------------------------------------
>
> Key: ISPN-1127
> URL: https://issues.jboss.org/browse/ISPN-1127
> Project: Infinispan
> Issue Type: Bug
> Components: Loaders and Stores
> Affects Versions: 4.2.1.FINAL, 5.0.0.CR3
> Reporter: Sergey Kondratyev
> Assignee: Manik Surtani
> Attachments: test-leak.tar.bz2
>
>
> CacheStoreInterceptor leaks org.infinispan.transaction.xa.GlobalTransaction objects. Amount of live org.infinispan.transaction.xa.GlobalTransaction objects is climbing during cache work.
> It is possible due to two reasons:
> 1. when cache is in async replication mode entries from both txStores and preparingTxs are never deleted (maybe like JBCACHE-1454 ?).
> 2. when cache is in sync mode (I tested on dist sync mode), entries from preparingTxs are not deleted in commit method (in case no exceptions raise).
> So preparingTxs are never cleaned at all. Problem with txStores appears only with async mode.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
More information about the infinispan-issues
mailing list