You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@activemq.apache.org by "ASF GitHub Bot (JIRA)" <ji...@apache.org> on 2017/04/12 13:41:41 UTC
[jira] [Commented] (ARTEMIS-1111) AMQP hangs if consumer is closed
during delivery
[ https://issues.apache.org/jira/browse/ARTEMIS-1111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15965868#comment-15965868 ]
ASF GitHub Bot commented on ARTEMIS-1111:
-----------------------------------------
GitHub user mtaylor opened a pull request:
https://github.com/apache/activemq-artemis/pull/1202
ARTEMIS-1111 Avoid deadlock on AMQP delivery during close
I wanted to avoid the possibility of adding more race/lock conditions by introducing new Executors/Threads when dealing with the situation described by ARTEMIS-1111. Instead I've upgraded the connection lock to a re-entrant lock. Instead of the consumer delivery thread blocking forever when the ProtonHandler lock is held by the close() event, it does a tryLock(), then checks to see if the link is closed. If so it will exit. If not it will retry.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/mtaylor/activemq-artemis ARTEMIS-1111
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/activemq-artemis/pull/1202.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #1202
----
commit 153053f8c71585bdfd66c353269a9e87aa1942f7
Author: Martyn Taylor <mt...@redhat.com>
Date: 2017-04-12T13:38:06Z
ARTEMIS-1111 Avoid deadlock on AMQP delivery during close
----
> AMQP hangs if consumer is closed during delivery
> ------------------------------------------------
>
> Key: ARTEMIS-1111
> URL: https://issues.apache.org/jira/browse/ARTEMIS-1111
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Reporter: Martyn Taylor
>
> There's a lock contention issue in the AMQP protocol. If a linkClose is processed whilst Artemis is delivering messages on the associated consumer, a deadlock scenario can occur. It will eventually timeout but the issue is still present.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)