hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tsz Wo (Nicholas), SZE (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-5131) Provide better handling of job status related apis during JT restart
Date Fri, 05 Apr 2013 16:49:15 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-5131?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13623783#comment-13623783
] 

Tsz Wo (Nicholas), SZE commented on MAPREDUCE-5131:
---------------------------------------------------

The defaultPolicy is to retry on JobTrackerNotYetInitializedException.  For SafeModeException,
defaultPolicy is equivalent to TRY_ONCE_THEN_FAIL since SafeModeException will be thrown as
a RemoteException.  So remoteExceptionToPolicyMap.put(SafeModeException.class, defaultPolicy)
is the same as remoteExceptionToPolicyMap.put(SafeModeException.class, TRY_ONCE_THEN_FAIL).

I think it is simpler to change RetryUtils.getDefaultRetryPolicy to support multiple exceptions.


                
> Provide better handling of job status related apis during JT restart
> --------------------------------------------------------------------
>
>                 Key: MAPREDUCE-5131
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5131
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>            Reporter: Arun C Murthy
>            Assignee: Arun C Murthy
>         Attachments: MAPREDUCE-5131.patch
>
>
> I've seen pig/hive applications bork during JT restart since they get NPEs - this is
due to fact that jobs are not really inited, but are submitted.

--
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