[
https://issues.apache.org/jira/browse/QPID-7357?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15382600#comment-15382600
]
ASF subversion and git services commented on QPID-7357:
-------------------------------------------------------
Commit 57e8b4a6bb2da5701804d988b875f335f64c64a5 in qpid-cpp's branch
refs/heads/master from [~aconway]
[ https://git-wip-us.apache.org/repos/asf?p=qpid-cpp.git;h=57e8b4a ]
QPID-7357: c++ HA Backup crash during re-connect in failover
The problem: A backup creates an UpdateTracker to delete non-existent queues at
the end of an update, in the UpdateTracker destructor. However if an update is
interrupted, it can leave a partially-used UpdateTracker behind. When the next
update starts this can incorrectly delete queues at a sensitive moment and cause
a crash, or could cause queues to be removed incorrectly.
The fix is to move queue deletion out of the destructor so it is only run at
the end of a successful update.
> c++ HA Backup crash during re-connect in failover
> -------------------------------------------------
>
> Key: QPID-7357
> URL: https://issues.apache.org/jira/browse/QPID-7357
> Project: Qpid
> Issue Type: Bug
> Components: C++ Broker, C++ Clustering
> Affects Versions: qpid-cpp-0.34
> Reporter: Alan Conway
> Assignee: Alan Conway
> Fix For: qpid-cpp-next
>
>
> A backup creates an UpdateTracker to delete non-existent queues at the end of
> an update, in the UpdateTracker destructor. However if an update is
> interrupted, it can leave a partially-used UpdateTracker behind. When the
> next update starts this can incorrectly delete queues at a sensitive moment
> and cause a crash.
> See also https://bugzilla.redhat.com/show_bug.cgi?id=1333767
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
