[
https://issues.apache.org/jira/browse/HDFS-1811?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13016192#comment-13016192
]
Allen Wittenauer commented on HDFS-1811:
----------------------------------------
If only we had a distributed configuration system....
I'm guessing this is actually for Federated branch? This doesn't seem
necessary for a single/dual NN config.
> Create scripts to decommission datanodes
> ----------------------------------------
>
> Key: HDFS-1811
> URL: https://issues.apache.org/jira/browse/HDFS-1811
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Owen O'Malley
> Assignee: Erik Steffl
>
> Create scripts to decommission datanodes:
> - distribute exclude file
> - input is location of exclude file
> - location on namenodes: hdfs getconf -excludeFile
> - list of namenodes: hdfs getconf -namenodes
> - scp excludes files to all namenodes
> - refresh namenodes
> - list of namenodes: hdfs getconf -namenodes
> - refresh namenodes: hdfs dfsadmin -refreshNodes
> Two scripts are needed because each of them might require different
> permissions.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
