[
https://issues.apache.org/jira/browse/HADOOP-10050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13797432#comment-13797432
]
Arpit Gupta commented on HADOOP-10050:
--------------------------------------
bq. The capacity-scheduler configs are removed from SingleCluster.apt.vm - any
reason for this?
I did not have to make any changes to be able to run an example.
We have yarn.scheduler.capacity.root.default.capacity=100 and
yarn.scheduler.capacity.root.queues=default.
The following 2 are not there in the configs that are part of the release
yarn.scheduler.capacity.root.capacity and
yarn.scheduler.capacity.root.unfunded.capacity
bq. Generated SingleCluster.html has "Set $$YARN_CONF_DIR" - mind fixing this
too?
Will fix this.
bq. yarn-daemon.sh is not meant to be in HADOOP_MAPRED_HOME.
Actually when i ran the hadoop-2.2.0 tar ball i did not have to set any of
these env vars everything was in ./sbin. I can make the change to to
HADOOP_YARN_HOME and HADOOP_MAPRED_HOME.
I cannot find any reference for HADOOP_PREFIX on
http://hadoop.apache.org/docs/r2.2.0/hadoop-project-dist/hadoop-common/SingleCluster.html
> Update single node and cluster install instructions to work with latest bits
> ----------------------------------------------------------------------------
>
> Key: HADOOP-10050
> URL: https://issues.apache.org/jira/browse/HADOOP-10050
> Project: Hadoop Common
> Issue Type: Bug
> Components: documentation
> Affects Versions: 2.2.0
> Reporter: Arpit Gupta
> Assignee: Arpit Gupta
> Priority: Minor
> Attachments: ClusterSetup.html, HADOOP-10050.patch, mapred-site.xml,
> SingleCluster.html, yarn-site.xml
>
>
> A few things i noticed
> 1. changes to yarn.nodemanager.aux-services
> 2. Set the framework to yarn in mapred-site.xml
> 3. Start the history server
> Also noticed no change to the capacity scheduler configs was needed.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
