[
https://issues.apache.org/jira/browse/MAPREDUCE-3054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13114801#comment-13114801
]
Vinod Kumar Vavilapalli commented on MAPREDUCE-3054:
----------------------------------------------------
bq. I dont think we can get away from not sending kill to the RM at all times.
I dont think we can rely on AM to be responsive and do the right work. Thoughts?
I agree with the general sentiment here. But the way I'd handle this is by
having different UI for graceful-kill and force-terminate: "bin/hadoop job
-kill" will do the usual graceful exit. If someone intends to do a "bin/hadoop
job -kill -force", they know the risks (of not getting history file,
notification etc).
In a way, we are arguing SIGTERM vs SIGKILL semantics here. Sure someone may
trap SIGTERM, but SIGTERM only promises a graceful exit if the process is
willing to. Otherwise, you go use SIGKILL. We don't need the framework to wrap
around SIGTERM and SIGKILL.
> Unable to kill submitted jobs
> -----------------------------
>
> Key: MAPREDUCE-3054
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3054
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.0
> Reporter: Siddharth Seth
> Assignee: Mahadev konar
> Priority: Blocker
> Fix For: 0.23.0
>
> Attachments: MAPREDUCE-3054.patch, MAPREDUCE-3054.patch,
> MAPREDUCE-3054.patch, MAPREDUCE-3054.patch
>
>
> Found by Philip Su
> The "mapred job -kill" command
> appears to succeed, but listing the jobs again shows that the job supposedly
> killed is still there.
> {code}
> mapred job -list
> Total jobs:2
> JobId State StartTime UserName Queue Priority
> SchedulingInfo
> job_1316203984216_0002 PREP 1316204924937 hadoopqa default NORMAL
> job_1316203984216_0001 PREP 1316204031206 hadoopqa default NORMAL
> mapred job -kill job_1316203984216_0002
> Killed job job_1316203984216_0002
> mapred job -list
> Total jobs:2
> JobId State StartTime UserName Queue Priority
> SchedulingInfo
> job_1316203984216_0002 PREP 1316204924937 hadoopqa default NORMAL
> job_1316203984216_0001 PREP 1316204031206 hadoopqa default NORMAL
> {code}
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
