[
https://issues.apache.org/jira/browse/MAPREDUCE-7366?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17433853#comment-17433853
]
Steve Loughran edited comment on MAPREDUCE-7366 at 10/25/21, 4:16 PM:
----------------------------------------------------------------------
closing as a wontfix as
* v2 is broken
* this doesn't work with v1 as v1 job commit assumes exclusive access to the
dest dir
* general fear of going near FileOutputCommitter
Proposed a patch in MAPREDUCE-7341
was (Author: [email protected]):
closing as a wontfix as
* v2 is broken
* this doesn't work with v1 as v1 job commit assumes exclusive access to the
dest dir
* general fear of going near FileOutputCommitter
Proposed a patch in MAPREDUCE-7366
> FileOutputCommitter Enable Concurent Writes
> --------------------------------------------
>
> Key: MAPREDUCE-7366
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-7366
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Reporter: ismail
> Priority: Major
> Labels: committers, easyfix, easytask, pull-request-available
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> is it possible to make `{{PENDING_DIR_NAME}}` configurable?
> That will enable concurrent writes to same location. current if two spark
> processes write same destination one of them is failing.
> current
> {code:java}
> public static final String PENDING_DIR_NAME = "_temporary";{code}
> new:
> {code:java}
> PENDING_DIR_NAME = conf.get("mapreduce.fileoutputcommitter.pending.dir",
> "_temporary");{code}
> here is custom commiter doing it:
> https://gist.github.com/ismailsimsek/33c55d8e1fcfc79160483c38a978edbd
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
