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 "Steve Loughran (JIRA)" <ji...@apache.org> on 2018/02/26 16:05:00 UTC
[jira] [Commented] (MAPREDUCE-7060) Cherry Pick PathOutputCommitter
class/factory to branch-3.0 & 2.10
[ https://issues.apache.org/jira/browse/MAPREDUCE-7060?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16377087#comment-16377087 ]
Steve Loughran commented on MAPREDUCE-7060:
-------------------------------------------
For branch-3 we just need to pull in the final changes of HADOOP-13786, excluding the s3a factory declaration in mapred-default
> Cherry Pick PathOutputCommitter class/factory to branch-3.0 & 2.10
> ------------------------------------------------------------------
>
> Key: MAPREDUCE-7060
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-7060
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Affects Versions: 3.0.0, 2.10.0
> Reporter: Steve Loughran
> Assignee: Steve Loughran
> Priority: Minor
> Attachments: MAPREDUCE-7060-branch-3.0-001.patch
>
>
> It's easier for downstream apps like Spark to pick up the new PathOutputCommitter superclass if it is there on 2.10+, even if the S3A committer isn't there.
> Adding the interface & binding stuff of HADOOP-6956 allows for third party committers to be deployed.
> I'm not proposing a backport of the HADOOP-13786 committer: that's Java 8, S3Guard, etc. Too traumatic. All I want here is to allow downstream code to be able to pick up the new interface and so be able to support it and other store committers when available
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: mapreduce-issues-help@hadoop.apache.org