[
https://issues.apache.org/jira/browse/HBASE-2068?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12795032#action_12795032
]
Lars George commented on HBASE-2068:
------------------------------------
Another inconsistency is that MasterMetrics uses a MetricsIntValue
{code}
/*
* Count of requests to the cluster since last call to metrics update
*/
private final MetricsIntValue cluster_requests =
new MetricsIntValue("cluster_requests", registry);
{code}
and the RegionsServerMetrics uses a MetricsRate class
{code}
/*
* Count of requests to the regionservers since last call to metrics update
*/
private final MetricsRate requests = new MetricsRate("requests");
{code}
I suggest after fixing MetricsRate to replace the MasterMetrics one with it as
well.
I am also wondering if it would have been better to call it MetricsIntRate to
indicate the size of the internal counter.
> MetricsRate is missing "registry" parameter
> -------------------------------------------
>
> Key: HBASE-2068
> URL: https://issues.apache.org/jira/browse/HBASE-2068
> Project: Hadoop HBase
> Issue Type: Bug
> Affects Versions: 0.20.2
> Reporter: Lars George
> Assignee: Lars George
> Priority: Minor
> Fix For: 0.20.3, 0.21.0
>
>
> I am trying to get the graphing going using Edward Capriolo's great JMX to
> Cacti [guide|http://www.jointhegrid.com/hadoop/]. I checked and I am missing
> the request rate in the JMX MBean:
> {noformat}
> # sh 0_20/regionserver/RegionServerStatistics.sh
> service:jmx:rmi:///jndi/rmi://foobar:10102/jmxrmi controlRole mypass
> hadoop:name=RegionServerStatistics,service=RegionServer
> storefiles:493 blockCacheFree:139771296 storefileIndexSizeMB:102
> memstoreSizeMB:0 stores:658 blockCacheCount:8400
> regions:83 blockCacheHitRatio:0 blockCacheSize:717478944
> atomicIncrementTimeNumOps:0 atomicIncrementTimeAvgTime:0
> atomicIncrementTimeMinTime:-1 atomicIncrementTimeMaxTime:0
> {noformat}
> I checked the code and the difference between requests and the other
> attributes is that MetricsRate does not register itself in the
> MetricsRegistry used by the dynamic MBean like for example the
> MetricsLongValue does.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
