[
https://issues.apache.org/jira/browse/ARTEMIS-1134?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15989413#comment-15989413
]
ASF GitHub Bot commented on ARTEMIS-1134:
-----------------------------------------
GitHub user tabish121 opened a pull request:
https://github.com/apache/activemq-artemis/pull/1236
ARTEMIS-1134 Close connection if error caught during event processing
If an error escapes into the event processing layer we close the
connection with an error condition to avoid the client becoming stuck on
waiting for a response from the broker and the broker side being in an
unknown state.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/tabish121/activemq-artemis ARTEMIS-1134
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/activemq-artemis/pull/1236.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 #1236
----
commit 4f9f5fc041f791eb2a829b2b65ce9dbb40f85116
Author: Timothy Bish <[email protected]>
Date: 2017-04-28T20:21:51Z
ARTEMIS-1134 Close connection if error caught during event processing
If an error escapes into the event processing layer we close the
connection with an error condition to avoid the client becoming stuck on
waiting for a response from the broker and the broker side being in an
unknown state.
----
> Internal errors in AMQP protocol handler can lead to stalled connection
> -----------------------------------------------------------------------
>
> Key: ARTEMIS-1134
> URL: https://issues.apache.org/jira/browse/ARTEMIS-1134
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Components: AMQP
> Affects Versions: 2.0.0
> Reporter: Timothy Bish
> Assignee: Timothy Bish
> Fix For: 2.next
>
>
> In the case where an exception slips through the AMQP event handling code and
> ends up caught in the event dispatching code the connection has an error
> condition set on it but is never closed, leaving the remote with no answer to
> the requested operation and the client must wait for an idle timeout if one
> is set before the connection is closed. The broker should remotely close the
> connection if an uncaught exception gets up into the event dispatcher so that
> the client is aware that something has gone wrong and requests / connections
> aren't left hanging.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
