[
https://issues.apache.org/jira/browse/HADOOP-2132?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jothi Padmanabhan updated HADOOP-2132:
--------------------------------------
Fix Version/s: 0.18.0
Status: Patch Available (was: Open)
The patch is available for review.
Changes:
1. Modified the Kill Method in JobInProgress class to check if the job is
either in the "RUNNING" or "PREP" state before killing it and moving it to the
failed state.
2. Modified "isSuccessful", "toString", and "is Complete" methods in JobClient
to forcefully update their job status (instead of using the cached status) to
return the most up-to-date status.
> Killing successfully completed jobs moves them to failed
> --------------------------------------------------------
>
> Key: HADOOP-2132
> URL: https://issues.apache.org/jira/browse/HADOOP-2132
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Environment: Hadoop cluster, after running one job
> Reporter: Srikanth Kakani
> Assignee: Jothi Padmanabhan
> Priority: Critical
> Fix For: 0.18.0
>
> Attachments: hadoop-2132.patch
>
>
> Once a job is in completed state, run hadoop job -kill <jobid>, it moves to
> failed queue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
