[ 
https://issues.apache.org/jira/browse/HBASE-3102?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12922787#action_12922787
 ] 

Gary Helmling commented on HBASE-3102:
--------------------------------------

Sorry, I mis-spoke above (that will teach me to reply to JIRA emails from a 
phone).  We already extend MetricsDynamicMBeanBase as 
org.apache.hadoop.hbase.metrics.MetricsMBeanBase because we already ran into 
the overly restrictive introspection issue.  For a new metrics class, just add 
it to MetricsMBeanBase.getAttribute() and you're done.  Aside from implementing 
the metrics class itself, of course, which Hadoop 0.20 unfortunately makes 
difficult.

For me the only issue is whether the configuration is obvious from an ops 
perspective.  It doesn't seem clear that the configuration would affect 
"compaction" here vs others, which was why I suggested the per-metric override.

That said this is a pretty simple change.  Can we just change the config to 
something like 

"hbase.regionserver.extendedrefresh" or 
"hbase.regionserver.compactionrefresh"

to make it a little clearer?

> Enhance HRegionServerMetrics 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
>
>
> 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.

Reply via email to