[
https://issues.apache.org/jira/browse/HADOOP-3204?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12587293#action_12587293
]
Devaraj Das commented on HADOOP-3204:
-------------------------------------
Some nits - the method getFailureCause should be called something else. While
the name sounds like it can tell us the failure cause whatsoever but in reality
you just depend on the mergeThrowable field. Actually at this point of time, it
might make sense to do something better - whenever you get a merge error (be it
ramfs or localfs), you can probably invoke umbilical.shuffleError and die.
> LocalFSMerger needs to catch throwable
> --------------------------------------
>
> Key: HADOOP-3204
> URL: https://issues.apache.org/jira/browse/HADOOP-3204
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Environment: Revision: 645751
> Reporter: Koji Noguchi
> Assignee: Amar Kamat
> Priority: Blocker
> Fix For: 0.17.0
>
> Attachments: HADOOP-3204-v1.patch
>
>
> I haven't used trunk(0.17) so I don't know what would happen when FSError is
> thrown within LocalFSMerger thread.
> Does it have the same problem as HADOOP-3154?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
