[jira] [Resolved] (SAMZA-1568) Handle ZkInterruptedException in zkclient.close.

["Jake Maes (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SAMZA-1568?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jake Maes resolved SAMZA-1568.
------------------------------
    Resolution: Fixed

PR merged and closed

> Handle ZkInterruptedException in zkclient.close.
> ------------------------------------------------
>
>                 Key: SAMZA-1568
>                 URL: https://issues.apache.org/jira/browse/SAMZA-1568
>             Project: Samza
>          Issue Type: Bug
>            Reporter: Shanthoosh Venkataraman
>            Assignee: Shanthoosh Venkataraman
>            Priority: Major
>
> When zookeeper session failures occur in a stream processor,   leaves the group(zkClient is closed) and joins the group again.
> The last step in that shutdown sequence is zkClient.close(). In some scenarios, it throws the following exception, 
> {code:java}
>     org.I0Itec.zkclient.exception.ZkInterruptedException: java.lang.InterruptedException
>     at org.I0Itec.zkclient.ZkClient.close(ZkClient.java:1278)
>     at org.apache.samza.zk.ZkControllerImpl.stop(ZkControllerImpl.java:92)
>     at org.apache.samza.zk.ZkJobCoordinator.stop(ZkJobCoordinator.java:141)
> {code}
> In existing implementation this is not handled, there by killing the stream processor.  The following codepath triggers this exception:
> {code:java}
> StreamProcessor.stop -> ZkJobCoordinator.stop() ->  zkController.stop() -> zkUtils.close
> {code}
> This exception causes the integration test to fail occasionally  and can cause LocalApplicationRunner.waitForFinish method call to block indefinitely(since this callback event success, updates the latch state required for waitForFinish to end).



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)