[
https://issues.apache.org/jira/browse/QPID-5366?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13827912#comment-13827912
]
ASF subversion and git services commented on QPID-5366:
-------------------------------------------------------
Commit 1543893 from [~aconway] in branch 'qpid/trunk'
[ https://svn.apache.org/r1543893 ]
QPID-5366: qpid segfaults in qpid::ha::BrokerReplicator::disconnected
Fix for a race condition: previously, BrokerReplicator created a separate
ConnectionObserver object to forward connection events to it. However the
Observers locking is such that it is possible for an event to arrive *after*
calling Observers::remove (Observers copies the pointers and delivers events
outside its lock.) This meant that it was possible for a call to
BrokerReplicator::disconnect to be made after the BrokerReplicator was deleted.
The fix is to combine BrokerReplicator and BrokerReplicator::ConnectionObserver
into a single object with one lifetime that will last until it is removed from
both the ExchangeRegistry and the ConnectionObservers.
> qpid segfaults in qpid::ha::BrokerReplicator::disconnected
> -----------------------------------------------------------
>
> Key: QPID-5366
> URL: https://issues.apache.org/jira/browse/QPID-5366
> Project: Qpid
> Issue Type: Bug
> Components: C++ Clustering
> Affects Versions: 0.24
> Reporter: Alan Conway
> Assignee: Alan Conway
>
> Relocating the qpidd-primary service while the cluster is under load was
> causing sporadic core dumps in BrokerReplicator::disconnected. For details
> see:
> https://bugzilla.redhat.com/show_bug.cgi?id=1030608
--
This message was sent by Atlassian JIRA
(v6.1#6144)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org
