[
https://issues.apache.org/jira/browse/HADOOP-3386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12596668#action_12596668
]
Arun C Murthy commented on HADOOP-3386:
---------------------------------------
Zheng, I'm not so sure about the problem you are trying to fix, though I'd
readily admit it's too hairy to keep all code-paths coherently in my head.
Just some pointers: TaskTracker.TaskInProgress.cleanup method is the one who
deletes local dirs. That call does not solely depend on KillJobAction as I
gather from this jira. It is called from TaskTracker.purgeTask and
TaskTracker.TaskInProgress.taskFinished too, so please be aware of those.
Overall, as I pointed out in HADOOP-3370, I'd be very happy to have you
implement what you propose - it is a very useful feature; I'm only asking you
to be clear about the direction of this jira...
> the job directory of a failed task may stay forever on a tasktracker node
> -------------------------------------------------------------------------
>
> Key: HADOOP-3386
> URL: https://issues.apache.org/jira/browse/HADOOP-3386
> Project: Hadoop Core
> Issue Type: Bug
> Reporter: Zheng Shao
>
> See https://issues.apache.org/jira/browse/HADOOP-3370 for details of the
> problem.
> A tasktracker only cleans out the job dir when the job tracker sends a
> "KILLJOB" action in the heartbeat response message.
> However, in a corner case, the job tracker will NOT send the "KILLJOB" action
> to the task tracker. The case is when there is only failed tasks of this job
> on this task tracker; no successful tasks of this job is on this task tracker.
> In this case, jobtracker.trackerToTaskMap will not contain an entry of this
> task tracker to any tasks of this job. As a result, the job tracker will not
> send a KILLJOB action to the task tracker.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
