hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nicolas Spiegelberg (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-3102) Enhance HBase rMetrics for Long-running Stats
Date Tue, 26 Oct 2010 22:03:21 GMT

    [ https://issues.apache.org/jira/browse/HBASE-3102?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12925168#action_12925168
] 

Nicolas Spiegelberg commented on HBASE-3102:
--------------------------------------------

@Gary: I started looking into the MetricsMBeanBase you suggested, but I was a little confused
on how exactly to hook up a new class.  I thought that I would make less programming mistakes
if I stuck with subclassing.  I agree the naming is kinda hacky, but of relatively trivial
importance.  And really, the naming fix should be made on the Hadoop side of things.  Create
some 'get/setSuffix()' functionality to MetricsTimeVaryingRate and use reflection to setSuffix("").
 The same argument could be made about data persistence, but persistence is not really optional
functionality to maintain useable metrics.  

I think the appropriate next step to completely resolve these loose ends is to file Hadoop
jiras for all these issues.  Then, we'll migrate away from using the PersistentMetricsTimeVaryingRate
over time.

> Enhance HBase rMetrics for Long-running Stats
> ---------------------------------------------
>
>                 Key: HBASE-3102
>                 URL: https://issues.apache.org/jira/browse/HBASE-3102
>             Project: HBase
>          Issue Type: Improvement
>          Components: regionserver
>            Reporter: Nicolas Spiegelberg
>            Assignee: Nicolas Spiegelberg
>             Fix For: 0.90.0
>
>         Attachments: HBASE-3102_90.patch
>
>
> Example: a useful metric to observe track is compaction count + duration. Since compactions
are long running and only happen infrequently, the avg/opcount stats are should be reset longer
than the default polling period (5 sec). In addition to 'hbase.period', we should allow a
different duration after which long-running metrics should expire. This would also fix our
existing metrics problem where min/max stats are never reset until the process is restarted/upgraded.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message