hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrew Wang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-10279) Improve the validation for tolerated volumes
Date Tue, 12 Apr 2016 23:56:25 GMT

    [ https://issues.apache.org/jira/browse/HDFS-10279?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15238282#comment-15238282
] 

Andrew Wang commented on HDFS-10279:
------------------------------------

Looks good, thanks for the patch! Only a few comments:

* Looks like we don't need the data dirs themselves, just the #. Thus can we store a count
instead?
* Let's use {{GenericTestUtils.assertExceptionContains}} to validate the DiskErrorException
in the test.

Also do you mind setting the affects and target version for this JIRA? It's good practice
when filing a new JIRA.

> Improve the validation for tolerated volumes
> --------------------------------------------
>
>                 Key: HDFS-10279
>                 URL: https://issues.apache.org/jira/browse/HDFS-10279
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>            Reporter: Lin Yiqun
>            Assignee: Lin Yiqun
>         Attachments: HDFS-10279.001.patch
>
>
> Now the misconfiguration for dfs.datanode.failed.volumes.tolerated are detected too late
and not easily be found. We can move the validation logic for tolerated volumes to a eariler
time that before datanode regists to namenode. And this will let us detect the misconfiguration
soon and easily.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message