hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sanjay Radia (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-2849) Use the MapReduce output committer
Date Tue, 13 Aug 2013 21:18:49 GMT

    [ https://issues.apache.org/jira/browse/HIVE-2849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13738846#comment-13738846
] 

Sanjay Radia commented on HIVE-2849:
------------------------------------

Hive also writes output to /tmp and renames to final output dir; this makes use of hive under
federation problematic since one often wants /tmp to be a separate volume and renames are
currently not supported across hdfs volumes.  Using the output committer will mean that the
temporary output is in the output dir's subtree and hence the rename works.
                
> Use the MapReduce output committer
> ----------------------------------
>
>                 Key: HIVE-2849
>                 URL: https://issues.apache.org/jira/browse/HIVE-2849
>             Project: Hive
>          Issue Type: Improvement
>            Reporter: Owen O'Malley
>            Assignee: Owen O'Malley
>
> Currently Hive implements its own output committers based on the size of the files using
regexes to parse out the task ids. The MapReduce output committers are very stable and use
a two phase commit from the JobTracker.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message