[
https://issues.apache.org/jira/browse/ARTEMIS-2200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16745487#comment-16745487
]
ASF subversion and git services commented on ARTEMIS-2200:
----------------------------------------------------------
Commit 0460bf25513c1af943e2c12358a8429567011bd3 in activemq-artemis's branch
refs/heads/2.6.x from Francesco Nigro
[ https://gitbox.apache.org/repos/asf?p=activemq-artemis.git;h=0460bf2 ]
ARTEMIS-2200 NPE while dropping/failing large messages on paging
Large messages pendingRecordID is not accessed atomically, leading
to races that would lead to records that cannot been found on the
journal for deletion: it would lead to cause NPE that won't clean
the pending tasks on the current OperationContextImpl.
Adding a cleanup on error of those tasks and avoiding the race
to happen by adding proper synchronization will both enforce
correct clean up when something bad happen and avoid NPE.
(cherry picked from commit 702f445205de953a343c1beef436fc8b984c7624)
> NPE while dropping/failing large messages on paging
> ---------------------------------------------------
>
> Key: ARTEMIS-2200
> URL: https://issues.apache.org/jira/browse/ARTEMIS-2200
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Affects Versions: 2.7.0, 2.6.3
> Reporter: Francesco Nigro
> Assignee: Francesco Nigro
> Priority: Minor
> Fix For: 2.7.0
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> Due to a race condition while handling deletion (ie confirmation) of large
> messages dropped/failed due to paging, the Artemis journal try to delete the
> same records twice, throwing NPE while doing it.
> In addition, given that the NPE is unexpected, the broker is not cleaning up
> the pending stores on the OperationContext, leading to some leaking
> OperationContextImpl.TaskHolder instances.
> It is happening with any protocols ie Core, OpenWire, AMQP.
> Are not known yet other bad effects beyond the NPE and the leak.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
