You are viewing a plain text version of this content. The canonical link for it is here.

Posted to hdfs-issues@hadoop.apache.org by "Allen Wittenauer (JIRA)" <ji...@apache.org> on 2012/11/10 14:25:14 UTC

[jira] [Commented] (HDFS-4169) Add per-disk latency metrics to DataNode

    [ https://issues.apache.org/jira/browse/HDFS-4169?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13494682#comment-13494682 ] 

Allen Wittenauer commented on HDFS-4169:
----------------------------------------

This is definitely a good thing and something we've needed for a long time now. We need to tread carefully if we decide to act on the metrics we get back however, as latency may occur just due to the basic multi-tenant nature of Hadoop.
                
> Add per-disk latency metrics to DataNode
> ----------------------------------------
>
>                 Key: HDFS-4169
>                 URL: https://issues.apache.org/jira/browse/HDFS-4169
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: data-node
>    Affects Versions: 3.0.0
>            Reporter: Todd Lipcon
>
> Currently, if one of the drives on the DataNode is slow, it's hard to determine what the issue is. This can happen due to a failing disk, bad controller, etc. It would be preferable to expose per-drive MXBeans (or tagged metrics) with latency statistics about how long reads/writes are taking.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira