Datanodes should decomission themselves on volume failure
---------------------------------------------------------
Key: HDFS-1847
URL: https://issues.apache.org/jira/browse/HDFS-1847
Project: Hadoop HDFS
Issue Type: Improvement
Components: data-node
Reporter: Eli Collins
Fix For: 0.23.0
A DN should decommission itself rather than shutdown whenever the configured
threshold of volume failures (dfs.datanode.failed.volumes.tolerated) has been
reached. This allows the namenode to distinguish between a datanode that might
not come back up soon (and will need replication) vs one that is just
restarting.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
