[
https://issues.apache.org/jira/browse/MAPREDUCE-4126?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Allen Wittenauer updated MAPREDUCE-4126:
----------------------------------------
Fix Version/s: (was: 1.0.0)
> This is a fix both for oracle support and managing DB splits in general.
> ------------------------------------------------------------------------
>
> Key: MAPREDUCE-4126
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-4126
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: performance
> Affects Versions: 1.0.0
> Environment: Redhat 6.1 x86_64
> Reporter: Joseph Doss
> Labels: hadoop, newbie, patch
> Attachments: hadoop.dbinputformat.query.split.patch.patch
>
>
> This is and addition to an older patch that addressed oracle support. We
> found that in large result sets, the map reduce was both splitting the
> results ( as expected ) and running a duplicate ( full result set ) on one
> node, resulting in a long runtime that should have been reduced by splitting.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
