[ https://issues.apache.org/jira/browse/HDFS-4239?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14212757#comment-14212757 ]
Yongjun Zhang commented on HDFS-4239: ------------------------------------- HI [~qwertymaniac], My bad that I did not notice your earlier comment {quote} I just noticed Steve's comment referring the same - should've gone through properly before spending google cycles. I feel HDFS-1362 implemented would solve half of this - and the other half would be to make the removals automatic. Right now the checkDiskError does not eject if its slow - as long as its succeed, which would have to be done via this JIRA I think. The re-add would be possible via HDFS-1362. {quote} until now. So we need to use the functionality provided by HDFS-1362 to automatically remove a sick disk. It seems the original goal of HDFS-4239 is the same as HDFS-1362 (right?), and we can create a new jira for automatically removing a sick disk? Thanks. > Means of telling the datanode to stop using a sick disk > ------------------------------------------------------- > > Key: HDFS-4239 > URL: https://issues.apache.org/jira/browse/HDFS-4239 > Project: Hadoop HDFS > Issue Type: Improvement > Reporter: stack > Assignee: Yongjun Zhang > Attachments: hdfs-4239.patch, hdfs-4239_v2.patch, hdfs-4239_v3.patch, > hdfs-4239_v4.patch, hdfs-4239_v5.patch > > > If a disk has been deemed 'sick' -- i.e. not dead but wounded, failing > occasionally, or just exhibiting high latency -- your choices are: > 1. Decommission the total datanode. If the datanode is carrying 6 or 12 > disks of data, especially on a cluster that is smallish -- 5 to 20 nodes -- > the rereplication of the downed datanode's data can be pretty disruptive, > especially if the cluster is doing low latency serving: e.g. hosting an hbase > cluster. > 2. Stop the datanode, unmount the bad disk, and restart the datanode (You > can't unmount the disk while it is in use). This latter is better in that > only the bad disk's data is rereplicated, not all datanode data. > Is it possible to do better, say, send the datanode a signal to tell it stop > using a disk an operator has designated 'bad'. This would be like option #2 > above minus the need to stop and restart the datanode. Ideally the disk > would become unmountable after a while. > Nice to have would be being able to tell the datanode to restart using a disk > after its been replaced. -- This message was sent by Atlassian JIRA (v6.3.4#6332)