[
https://issues.apache.org/jira/browse/SPARK-1064?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sandy Ryza reassigned SPARK-1064:
---------------------------------
Assignee: Sandy Ryza (was: Sandy Pérez González)
> Make it possible to use cluster's Hadoop jars when running against YARN
> -----------------------------------------------------------------------
>
> Key: SPARK-1064
> URL: https://issues.apache.org/jira/browse/SPARK-1064
> Project: Spark
> Issue Type: Sub-task
> Components: YARN
> Reporter: Sandy Pérez González
> Assignee: Sandy Ryza
> Fix For: 1.0.0
>
>
> YARN applications like MapReduce and Tez rely on the cluster's Hadoop jars
> instead of distributing their own.
> This has a couple advantages
> * Avoids sending a bunch of bits to every node for each app
> * Only a single version of Hadoop can be running on a cluster at one time,
> simplifying debugging
> * Easier to upgrade and apply patched versions of Hadoop
--
This message was sent by Atlassian JIRA
(v6.2#6252)
