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

Posted to mapreduce-issues@hadoop.apache.org by "Jason Lowe (JIRA)" <ji...@apache.org> on 2013/06/06 18:21:20 UTC

[jira] [Updated] (MAPREDUCE-5307) Failure to launch a task attempt leads to missing failed attempt

     [ https://issues.apache.org/jira/browse/MAPREDUCE-5307?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jason Lowe updated MAPREDUCE-5307:
----------------------------------

    Summary: Failure to launch a task attempt leads to missing failed attempt  (was: Failure to launch a task leads to missing failed task)
    
> Failure to launch a task attempt leads to missing failed attempt
> ----------------------------------------------------------------
>
>                 Key: MAPREDUCE-5307
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5307
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: mr-am, mrv2
>    Affects Versions: 0.23.7, 2.0.4-alpha
>            Reporter: Jason Lowe
>
> When the AM tries to connect to a node to launch a task attempt but the launcher encounters an error (e.g.: RPC timeout, no connectivity to node, etc.), the error will count as a failed task attempt.  However the task attempt will not appear in the job history leading to confusing situations.
> For example, one job that had particular difficulty connecting to nodes had a task fail four attempts.  Attempts 0, 1, and 3 timed out trying to launch the task, and attempt 2 failed for other reasons.  The job diagnostics were "Task task_1368637921738_2222198_m_000396 failed 1 times" and there was only one task attempt listed for task_1368637921738_2222198_m_000396,  which was attempt_1368637921738_2222198_m_000396_2.

--
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