hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Koji Noguchi (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-271) Metrics on Secondary namenode's activity
Date Wed, 22 Jun 2011 16:16:47 GMT

    [ https://issues.apache.org/jira/browse/HDFS-271?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13053328#comment-13053328
] 

Koji Noguchi commented on HDFS-271:
-----------------------------------

bq. Hey Koji, does the work done in HADOOP-3741 meet your requirements?

I'm afraid http status page won't help.  
If I remember from 2 years back, at that time I think we saw some slowness on the primary
namenode when fsimage was being copied back from the secondary namenode.

I wanted a metrics showing secondary namenode activity so that I have a way to look back and
correlate those slowness.  Now to think of it, what I really need is probably a 'checkpoint
activity' on the primary namenode instead of the secondary.


> Metrics on Secondary namenode's activity
> ----------------------------------------
>
>                 Key: HDFS-271
>                 URL: https://issues.apache.org/jira/browse/HDFS-271
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>            Reporter: Koji Noguchi
>            Priority: Minor
>
> To monitor secondary namenode's activiely, we currently rely on 'delete' metrics when
it replays the edits.
> (HADOOP-1495) 
> Requesting for new metrics that shows secondary namenode's activity.
> Maybe the size of the fsimage/edits being read and written.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message