Asaf Mesika 2013-06-05, 18:31
Otis Gospodnetic 2013-06-09, 05:55
-Re: Missing metrics in any context inheriting from AbstractMetricsContext
The first I did was to use JConsole to see what is available.
The thing that surprised me is the very small number of metrics exposed by
HBase Master on the "hbase" namespace.
When I hooked up "rpc" namespace and also did the Hadoop-metrics.properties
on the region server I finally saw a large amount of metrics.
So my mistake was enabling the FileTimstamp context class only on the
"hbase" namespace, and only on the Master.
On Sunday, June 9, 2013, Otis Gospodnetic wrote:
> I don't have a direct answer to your question, but if getting metrics
> from JMX is an option, everything should be in there. You can use a
> tool like https://github.com/sematext/jmxc to quickly get a JMX dump
> and see what's in there.
> HBase / Hadoop Performance Monitoring - http://sematext.com/spm/index.html
> > Hi,
> > I've tried using a GraphiteContext I've found in GitHub, but I
> discovered I
> > only get a very partial list of metrics.
> > I've switched to using TimeStampingFileContext and I get the same list:
> > 2013-06-05T18:24:59 hbase.master:
> > Master=a172-25-98-73.deploy.acme.com,60000,1370455069499,
> > hostName=a172-25-98-73.deploy.acme.com, cluster_requests=0.0,
> > splitSize_avg_time=0, splitSize_num_ops=0, splitTime_avg_time=0,
> > splitTime_num_ops=0
> > Does someone has any idea where are the rest of the metrics under "hbase"
> > context name?
> > This is my hadoop-metrics.properties:
> > hbase.class=org.apache.hadoop.hbase.metrics.file.TimeStampingFileContext
> > hbase.fileName=/tmp/metrics_hbase.log
> > hbase.period=10
> > # HBase-specific configuration to reset long-running stats (e.g.
> > compactions)
> > # If this variable is left out, then the default is no expiration.
> > hbase.extendedperiod = 3600
> > jvm.class=org.apache.hadoop.metrics.spi.NullContextWithUpdateThread
> > jvm.period=10
> > rpc.class=org.apache.hadoop.metrics.spi.NullContextWithUpdateThread
> > rpc.period=10