[
https://issues.apache.org/jira/browse/MAPREDUCE-1119?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12776896#action_12776896
]
Vinod K V commented on MAPREDUCE-1119:
--------------------------------------
We can avoid the uncertainity with dumping stack in case of Child exception by
directly calling {{taskController.dumpTaskStack(context)}} from inside
{{TaskTracker.markUnresponsiveTasks()}} immediately before
{{purgeTask(tip,true)}}. This will create the dump only when it is absolutely
needed.
Would that work? To construct the context, you may need a bridging method
inside {{JvmManager}} which can itself call
{{taskController.dumpTaskStack(context)}}.
That will make the patch a lot simpler, and will avoid many other avoidable
changes to {{JmvManager}}/{{TaskController}}. If you wish, I can upload a
demonstrating patch. Thoughts?
> When tasks fail to report status, show tasks's stack dump before killing
> ------------------------------------------------------------------------
>
> Key: MAPREDUCE-1119
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1119
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: tasktracker
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Aaron Kimball
> Attachments: MAPREDUCE-1119.2.patch, MAPREDUCE-1119.3.patch,
> MAPREDUCE-1119.4.patch, MAPREDUCE-1119.patch
>
>
> When the TT kills tasks that haven't reported status, it should somehow
> gather a stack dump for the task. This could be done either by sending a
> SIGQUIT (so the dump ends up in stdout) or perhaps something like JDI to
> gather the stack directly from Java. This may be somewhat tricky since the
> child may be running as another user (so the SIGQUIT would have to go through
> LinuxTaskController). This feature would make debugging these kinds of
> failures much easier, especially if we could somehow get it into the
> TaskDiagnostic message
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
