Closed mjcollin closed 6 years ago
Looks like rebalancing is manual for HDFS:
Trying:
hdfs balancer -threshold 25
to start. Will likely need to make this a cron job on a name node.
Looks like HDFS is happy now. Close this issue?
After recent fun w/ name node not syncing to secondary, I'm a bit cautious about cron'ing maintenance. Should not get out of balance too quickly in the future unless we do a lot of writes and deletes.
mesos01 has a full disk with the rest of the cluster only being about half full:
http://mesos02.acis.ufl.edu:50070/dfshealth.html#tab-datanode