helix-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "dafu (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HELIX-121) Possible race condition in ZkHelixManager.disconnect() during zk session expiry
Date Thu, 06 Jun 2013 17:39:21 GMT

    [ https://issues.apache.org/jira/browse/HELIX-121?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13677285#comment-13677285
] 

dafu commented on HELIX-121:
----------------------------

yes. this is the root case that we have multiple helix manager instances when zk session expiry
happens.
                
> Possible race condition in ZkHelixManager.disconnect() during zk session expiry
> -------------------------------------------------------------------------------
>
>                 Key: HELIX-121
>                 URL: https://issues.apache.org/jira/browse/HELIX-121
>             Project: Apache Helix
>          Issue Type: Bug
>            Reporter: dafu
>
> There is race condition in ZkHelixManager.disconnect() during zk session expiry. If someone
calls ZkHelixManager.disconnect() when zk session expiry happens, the real shutdown work may
be skipped:
> public void disconnect()
>   {
>     if (!isConnected())
>     {
>       logger.error("ClusterManager " + _instanceName + " already disconnected");
>       return;
>     }
>     disconnectInternal();
>   }
> And when zk connection comes back again, the zk manager instance will come alive again
even though the instance might be considered dead already.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message