Improvements to Doc: Hadoop MapReduce Next Generation - Cluster Setup
---------------------------------------------------------------------
Key: HADOOP-7799
URL: https://issues.apache.org/jira/browse/HADOOP-7799
Project: Hadoop Common
Issue Type: Bug
Components: documentation
Affects Versions: 0.23.0
Reporter: Eric Payne
Assignee: Arun C Murthy
- In section +Prerequisites+
-- Please add a link to download locaion of 0.23 release tar-ball.
- In section +Installation+
-- It would be good to have more details about where they should "untar" the
image and what the directory structure should look like. I can provide my notes
on the detailed steps I took to successfully install.
- In the section +Running Hadoop in Non-Secure Mode+
-- Can you add that templaces for the site and environment files can be found
in {{./conf/}}, {{./share/hadoop/common/templates/conf/}}, and
{{./share/hadoop/hdfs/templates/conf/}}
-- Also, it might be good to add a link to a sample configs that contain the
bare minumum to start a cluster.
-- The +conf/hdfs-site.xml+ section references the {{dfs.datanode.data.dir}}
property. Should we also include a reference to the
{{dfs.datanode.data.dir.perm}} property as well?
-- The +Hadoop Rack Awareness+ section references
{{topology.node.switch.mapping.impl}}, but I think this is deprecated. I think
the new one is {{net.topology.node.switch.mapping.impl}}. Also,
{{topology.script.file.name}} seems to be deprecated in favor of
{{net.topology.script.file.name}}.
- In section +Operating the Hadoop Cluster+
-- The HDFS format command should have the {{-clusterid}} parameter:
--- {{$HADOOP_PREFIX_HOME/bin/hdfs namenode -format -clusterid <cluster_name>}}
-- The command to start the namenode is incorrect. It should either be:
1. {{$HADOOP_PREFIX_HOME/bin/hdfs --config $HADOOP_CONF_DIR namenode &}} ##
without the {{start}}
or
2. {{$HADOOP_PREFIX/sbin/hadoop-daemon.sh --config $HADOOP_CONF_DIR start
namenode}}
I prefer the second, since you can also use that to stop the daemon.
-- The same for the command to start the datanode
-- The same for the command to start the resourcemanager, historyserver, and
nodemanagers, except that it should be {{yarn-daemon.sh}}
- Hadoop Shutdown:
-- If stopping the daemons is required, then {{hadoop-daemon.sh}} and
{{yarn-daemon.sh}} should be used to both start and stop.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
