You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@activemq.apache.org by "Guy Veraghtert (JIRA)" <ji...@apache.org> on 2008/09/29 16:10:53 UTC

[jira] Created: (AMQ-1958) JDBC master/slave deadlock when connection is lost

JDBC master/slave deadlock when connection is lost
--------------------------------------------------

                 Key: AMQ-1958
                 URL: https://issues.apache.org/activemq/browse/AMQ-1958
             Project: ActiveMQ
          Issue Type: Bug
          Components: Broker, Message Store
    Affects Versions: 5.2.0, 5.1.0, 5.0.0, 4.1.2
         Environment: oracle 10, mysql 5
            Reporter: Guy Veraghtert


In a pure JDBC failover scenario with one master and one slave: when the master loses its network connection to the database, the lock in the database will not be released. As such the slave will not know that the master isn't able to perform its work and it will continue to attempt to acquire a lock (which will never be released), when the master is restarted (after its connection is restored), it will become slave too, ending up with two slaves.

This behavior was encountered on Oracle 10 and Mysql 5.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

[jira] Updated: (AMQ-1958) JDBC master/slave deadlock when connection is lost

Posted by "Rob Davies (JIRA)" <ji...@apache.org>.

     [ https://issues.apache.org/activemq/browse/AMQ-1958?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Rob Davies updated AMQ-1958:
----------------------------

    Fix Version/s: 5.4.0

> JDBC master/slave deadlock when connection is lost
> --------------------------------------------------
>
>                 Key: AMQ-1958
>                 URL: https://issues.apache.org/activemq/browse/AMQ-1958
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Broker, Message Store
>    Affects Versions: 4.1.2, 5.0.0, 5.1.0, 5.2.0
>         Environment: oracle 10, mysql 5
>            Reporter: Guy Veraghtert
>             Fix For: 5.4.0
>
>
> In a pure JDBC failover scenario with one master and one slave: when the master loses its network connection to the database, the lock in the database will not be released. As such the slave will not know that the master isn't able to perform its work and it will continue to attempt to acquire a lock (which will never be released), when the master is restarted (after its connection is restored), it will become slave too, ending up with two slaves.
> This behavior was encountered on Oracle 10 and Mysql 5.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.