You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-issues@hadoop.apache.org by "Bhallamudi Venkata Siva Kamesh (Updated) (JIRA)" <ji...@apache.org> on 2012/02/20 09:19:43 UTC
[jira] [Updated] (MAPREDUCE-3471) Provide NullOutputCommitter
[ https://issues.apache.org/jira/browse/MAPREDUCE-3471?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Bhallamudi Venkata Siva Kamesh updated MAPREDUCE-3471:
------------------------------------------------------
Attachment: MAPREDUCE-3471.patch
I encountered the importance of *NullOutputCommitter*, when I was debugging [MAPREDUCE-3130|https://issues.apache.org/jira/browse/MAPREDUCE-3130]. I fixed this by setting *OutputCommitter* class of FileOutputFormat as *NullOutputCommitter*. Again for setting this, we need to fix [MAPREDUCE-2493|https://issues.apache.org/jira/browse/MAPREDUCE-2493].
Also, OC implementation of NOF returns true for isRecoverySupported(). Why do we need this?
Providing patch, thinking it may be useful.
> Provide NullOutputCommitter
> ---------------------------
>
> Key: MAPREDUCE-3471
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3471
> Project: Hadoop Map/Reduce
> Issue Type: New Feature
> Reporter: Bhallamudi Venkata Siva Kamesh
> Attachments: MAPREDUCE-3471.patch
>
>
> Hadoop by default provides only FileOutputCommitter and the same has been used internally except in the NOF.
> There are cases where using FOC is not much appropriate.
> Example
> DBOutputFormat instantiates FOC, though it does nothing.
> I think using NOC will be much more appropriate here.
> And also it can be used along with other OFs, where taking special care of the task/job output directories may not be required.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira