Details
-
New Feature
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
hot swappable disks
-
diskMaintenanceMode
-
Patch
Description
Like putting datanodes into maintenance mode, I don't want to have to reduce capacity to swap out failing drives in a hot swappable environment. I'd like to have a list of datanodes and storageIDs or datanode and mountpoints in a file that can act like the nodes in mainenance mode. This can ensure that blocks are replicated correctly and all replicas that exist on failing but not failed disks can be read off and replicated while the volumes are read only in hdfs land.