[
https://issues.apache.org/jira/browse/HADOOP-4361?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12637985#action_12637985
]
Vinod K V commented on HADOOP-4361:
-----------------------------------
Else where in HADOOP-4035, I needed to check if a job is being killed to avoid
duplicate kills. For this I had already introduced a boolean killInProgress
which is set in JobInProgress.terminate(). We'll need to sync up depending on
which JIRA goes in first.
> Corner cases in killJob from command line
> -----------------------------------------
>
> Key: HADOOP-4361
> URL: https://issues.apache.org/jira/browse/HADOOP-4361
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.19.0
> Reporter: Amareshwari Sriramadasu
> Assignee: Amareshwari Sriramadasu
> Priority: Blocker
> Fix For: 0.19.0
>
>
> If the job has occupied all the slots (maps and reduces), and there is kill
> from command line to kill the job, we have to make sure the cleanup task is
> eventually launched. Now there is no slot free until one of the attempts
> completes/failed.
> If there is kill from commandline during job setup, setup task has to be
> killed, before launching cleanup task.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
