[
https://issues.apache.org/jira/browse/HADOOP-15493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16513069#comment-16513069
]
Daryn Sharp commented on HADOOP-15493:
--------------------------------------
I don't think this disk-is-writable check should be in common. It only invites
(mis)use which makes it much easier to unwittingly cripple a cluster.
As we've seen, it's dangerous to assume that an IOE means a failed disk. Now
we are trying to whitelist a full disk but it's not straightforward. What
other transient IOEs may occur, perhaps only under load?
I think we have to rely on the system to detect a failed controller/drive.
Maybe we should just attempt to provoke the disk to go read-only. Have the DN
periodically write a file to its storages every n-many mins – but take _no_
action upon failure. Instead rely on the normal disk check to subsequently
discover the disk is read-only.
> DiskChecker should handle disk full situation
> ---------------------------------------------
>
> Key: HADOOP-15493
> URL: https://issues.apache.org/jira/browse/HADOOP-15493
> Project: Hadoop Common
> Issue Type: Improvement
> Reporter: Arpit Agarwal
> Assignee: Arpit Agarwal
> Priority: Critical
> Attachments: HADOOP-15493.01.patch, HADOOP-15493.02.patch
>
>
> DiskChecker#checkDirWithDiskIo creates a file to verify that the disk is
> writable.
> However check should not fail when file creation fails due to disk being
> full. This avoids marking full disks as _failed_.
> Reported by [~kihwal] and [~daryn] in HADOOP-15450.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
