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 subversion and git services (JIRA)" <ji...@apache.org> on 2016/10/24 11:52:58 UTC

[jira] [Commented] (AMQ-6476) Enabling durable sync can deadlock a network bridge on start up

    [ https://issues.apache.org/jira/browse/AMQ-6476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15601751#comment-15601751 ] 

ASF subversion and git services commented on AMQ-6476:
------------------------------------------------------

Commit 2ea6b00eef2efc779783314c4e0c220c774997b5 in activemq's branch refs/heads/activemq-5.14.x from [~cshannon]
[ https://git-wip-us.apache.org/repos/asf?p=activemq.git;h=2ea6b00 ]

https://issues.apache.org/jira/browse/AMQ-6476

Moving BrokerSubscriptionInfo processing into a new thread to prevent a
deadlock of the network bridge on startup

(cherry picked from commit b9cb02ae54a27e6c2f3ff934a1f01b88b7af2f11)


> Enabling durable sync can deadlock a network bridge on start up
> ---------------------------------------------------------------
>
>                 Key: AMQ-6476
>                 URL: https://issues.apache.org/jira/browse/AMQ-6476
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Broker
>    Affects Versions: 5.14.1
>            Reporter: Christopher L. Shannon
>            Assignee: Christopher L. Shannon
>
> There is an issue with the processing of the new BrokerSubscriptionInfo command that is used for sync durables over a network bridge that can cause the brokers to get stuck starting up.  Depending on the timing of the BrokerSubscriptionInfo being sent, each broker can get stuck waiting to finish creating static destinations and not be able to finish because it's waiting for the other broker to respond.  The root of the problem is the new countDown latch that was added when processing the new BrokerSubscriptionInfo object.  The fix for this to prevent a deadlock is to move that processing into a new thread so the transport doesn't get blocked and can continue to process other commands.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)