[jira] [Commented] (YARN-529) MR app master clean staging dir when reboot command sent from RM while the MR job succeeded

["jian he (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/YARN-529?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13620092#comment-13620092 ] 

jian he commented on YARN-529:
------------------------------

several solutions:
1. Let RM accept old attempts. In current case, RM will raise exception because unrecognized attempts and think the job unsuccessful
2. Only clean staging dir after AM successfully unregister with RM. We can use a flag to indicate or modify state machine when receive JOB_AM_REBOOT, transition from SUCCEEDED to REBOOT. The potential problem is that, when job transition to SUCCEEDED state, some job success metrics stuff has already been triggered.
                
> MR app master clean staging dir when reboot command sent from RM while the MR job succeeded
> -------------------------------------------------------------------------------------------
>
>                 Key: YARN-529
>                 URL: https://issues.apache.org/jira/browse/YARN-529
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: resourcemanager
>            Reporter: jian he
>            Assignee: jian he
>
> MR app master will clean staging dir, if the job is already succeeded and asked to reboot. RM will consider this job unsuccessful and launch further attempts, further attempts will fail because staging dir is cleaned

--
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