[
https://issues.apache.org/jira/browse/HADOOP-4236?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12639745#action_12639745
]
Sharad Agarwal commented on HADOOP-4236:
----------------------------------------
bq. Sameer has a good point that since the schedulers may delay initializing
jobs for an arbitrary length of time, we should probably kill jobs before they
get to init.
the current patch is already doing it. As far I understand, Arun and Sameer
have suggested to move those flags to a class.
bq. It would be far better if we could always make the kill method do the kill
and block if it is in the middle of initialization.
Agreed, that would be ideal and cleaner as well (just make kill as
synchronized). But we are trying to deal with the problem of jobtracker locking
- see HADOOP-3864. Currently during the kill job, the jobtracker is locked and
we can't afford to block the kill job if its in the middle of initTasks().
HADOOP-869 is seemingly becoming more important.
> JobTracker.killJob() fails to kill a job if the job is not yet initialized
> --------------------------------------------------------------------------
>
> Key: HADOOP-4236
> URL: https://issues.apache.org/jira/browse/HADOOP-4236
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.19.0
> Reporter: Amar Kamat
> Assignee: Sharad Agarwal
> Priority: Blocker
> Fix For: 0.19.0
>
> Attachments: 4236_v1.patch, 4236_v2.patch, 4236_v3.patch
>
>
> HADOOP-3864 made the following changes to {{JobTracker.killJob()}}
> {code}
> public synchronized void killJob(JobID jobid) {
> JobInProgress job = jobs.get(jobid);
> - job.kill();
> + if (job.inited()) {
> + job.kill();
> + }
> }
> {code}
> This is a bug as a job will not get killed if its not yet initialized.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
