[
https://issues.apache.org/jira/browse/MAPREDUCE-279?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12998654#comment-12998654
]
Tom White commented on MAPREDUCE-279:
-------------------------------------
bq. Hadoop needs to become more modular internally
+1 An easy way to achieve this here would be to put the resource manager code
and new MapReduce ApplicationMaster code into separate source trees under
mapreduce. This will help enforce dependencies from the beginning: MR2 depends
on MR1 and RM, but RM doesn't depend on anything else (except common for RPC?).
Going further, the work to separate out the API and libraries from the
implementation should help this effort too, since it will involve removing hard
dependencies on the jobtracker from the API classes (see MAPREDUCE-1478,
MAPREDUCE-1638).
> Map-Reduce 2.0
> --------------
>
> Key: MAPREDUCE-279
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-279
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: jobtracker, tasktracker
> Reporter: Arun C Murthy
> Assignee: Arun C Murthy
> Fix For: 0.23.0
>
>
> Re-factor MapReduce into a generic resource scheduler and a per-job,
> user-defined component that manages the application execution.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
