[
https://issues.apache.org/jira/browse/HADOOP-11353?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14235983#comment-14235983
]
Allen Wittenauer commented on HADOOP-11353:
-------------------------------------------
I'm expecting more people to use it to replace HADOOP_CLIENT_OPTS than anything
else.
Pointing to a different cluster sort of works; the java portions of
HADOOP_CONF_DIR and the paths used for classpath calculations are used *after*
.hadooprc gets processed. So theoretically, you could do something fun like:
{code}
/sw/hadoop.${CLUSTER}
/etc/hadoop -> dir that only contains a generic hadoop-env.sh
/etc/hadoop.mycoolcluster
.hadooprc:
HADOOP_CONF_DIR=/etc/hadoop.${CLUSTER}
HADOOP_COMMON_HOME=/sw/hadoop.${CLUSTER}
HADOOP_HDFS_HOME=/sw/hadoop.${CLUSTER}
HADOOP_YARN_HOME=/sw/hadoop.${CLUSTER}
HADOOP_MAPRED_HOME=/sw/hadoop.${CLUSTER}
$ CLUSTER=mycoolcluster yarn jar myjob.jar
{code}
For documentation, do think it should be in the CommandsManual? I've got a
bunch of stuff to document, actually... :)
> Add support for .hadooprc
> -------------------------
>
> Key: HADOOP-11353
> URL: https://issues.apache.org/jira/browse/HADOOP-11353
> Project: Hadoop Common
> Issue Type: New Feature
> Components: scripts
> Affects Versions: 3.0.0
> Reporter: Allen Wittenauer
> Assignee: Allen Wittenauer
> Labels: scripts
> Attachments: HADOOP-11353.patch
>
>
> The system should be able to read in user-defined env vars from ~/.hadooprc.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
