[
https://issues.apache.org/jira/browse/BIGTOP-1382?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14076437#comment-14076437
]
jay vyas commented on BIGTOP-1382:
----------------------------------
Hi jon ! You DEFINETLY do NOT want to put in that change to mapreduce_shuffle.
The reason is that for 0.8.0 (head) release, the parameter is correct.
You are deploying *0.7.0*, where you *manually* change that parameter, but we
dont want to bake that change into the puppet recipes.
Makes sense? otherwise, I like adding yarn and mapred components.
Please revert the change to the xml file, and maybe replace it with a comment
specifying that {{for 0.7.0, this should be mapreduce.shuffle, for 0.8+,
mapreduce_shuffle}} .
> Hadoop Component Additions and yarn-site.xml change
> ---------------------------------------------------
>
> Key: BIGTOP-1382
> URL: https://issues.apache.org/jira/browse/BIGTOP-1382
> Project: Bigtop
> Issue Type: Bug
> Components: VM
> Affects Versions: backlog
> Reporter: Jon Cope
> Priority: Trivial
> Labels: patch
> Fix For: backlog
>
> Attachments: BIGTOP-1382.patch
>
> Original Estimate: 0h
> Remaining Estimate: 0h
>
> A couple quick fixes to smooth installation of hadoop using
> /vm/vagrant-puppet/startup.sh.
> * added yarn and mapred-app to site.csv component list in
> /vagrant-puppet/provision.sh
> * changed yarn-site.xml property: yarn.nodemanager.aux.services =
> mapreduce_shuffle to mapreduce.shuffle; _see step 6:_ [Apache Wiki: How to
> Install Bigtop Hadoop with puppet |
> https://cwiki.apache.org/confluence/display/BIGTOP/How+to+install+BigTop+0.7.0+hadoop+on+CentOS+with+puppet]
--
This message was sent by Atlassian JIRA
(v6.2#6252)
