[
https://issues.apache.org/jira/browse/MAPREDUCE-430?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12747514#action_12747514
]
Arun C Murthy commented on MAPREDUCE-430:
-----------------------------------------
Aaron - I apologize for being unclear and assuming people got the context.
The distinction mainly arises from the desire to weed out node/hardware
failures as opposed to application errors. The the distinction is rooted in the
desire to treat node/hardware errors (disk, corrupt RAM/NIC etc.) differently
so as to quickly and accurately penalize the node (e.g. blacklist the
tasktracker). Currently we use _all_ task failures uniformly to penalize the
tasktracker... clearly penalizing a tasktracker for an application error which
results in an OOM (for e.g.) is injudicious. Does that make sense?
> Task stuck in cleanup with OutOfMemoryErrors
> --------------------------------------------
>
> Key: MAPREDUCE-430
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-430
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Amareshwari Sriramadasu
> Assignee: Amar Kamat
> Fix For: 0.20.1
>
> Attachments: MAPREDUCE-430-v1.11.patch,
> MAPREDUCE-430-v1.12-branch-0.20.patch, MAPREDUCE-430-v1.12.patch,
> MAPREDUCE-430-v1.6-branch-0.20.patch, MAPREDUCE-430-v1.6.patch,
> MAPREDUCE-430-v1.7.patch, MAPREDUCE-430-v1.8.patch
>
>
> Obesrved a task with OutOfMemory error, stuck in cleanup.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
