[
https://issues.apache.org/jira/browse/MAPREDUCE-1005?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12758142#action_12758142
]
Iyappan Srinivasan commented on MAPREDUCE-1005:
-----------------------------------------------
It can copy to ${dist.dir}/contrib/${contrib.project}/. In addition to copying
here, why cant it also put the related jar file under "lib" directory too.
Right now this copying from contrib directroy to lib directory is happening
manually, when building release tarball.
> why should not capacity sheduler jar not copied under lib during build
> ----------------------------------------------------------------------
>
> Key: MAPREDUCE-1005
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1005
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: build
> Affects Versions: 0.21.0
> Reporter: Iyappan Srinivasan
>
> Right now, the capacity-scheduler.jar is not getting copied under
> trunk/build/hadoop-mapred-0.21.0-dev/lib/. It stays in
> trunk/build/hadoop-mapred-0.21.0-dev/contrib/capacity-scheduler/hadoop-0.21.0-dev-capacity-scheduler.jar.
> Every time somebody tries to build the trunk, they have to copy this jar
> from contrib to lib manually.
> During any kind of build (ant binary, ant bin-package, ant package) , this
> jar file should also get copied to lib. If a person uses default or fair
> sheduler, he will not use it, but still there is no harm of it being in lib.
> If he uses capacity scheduler, then it gets used automatically. This manual
> copy should be avoided.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
