[ https://issues.apache.org/jira/browse/HBASE-14423?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14744434#comment-14744434 ]
stack commented on HBASE-14423: ------------------------------- So, even with retries I can get an odd failure of this test. Interesting is that with the extra info dumped by this patch, I see we are adding our bean in the correct place. We fail with: 2015-09-14 12:53:25,174 WARN [Time-limited test] hbase.TestStochasticBalancerJmxMetrics(229): Failed to get bean!!! Hadoop:service=HBase,name=Master,sub=Balancer javax.management.InstanceNotFoundException: Hadoop:service=HBase,name=Master,sub=Balancer ... but then after a bit of cycling I see this show up: Class Name: Master,sub=Balancer Object Name: Hadoop:service=DataNode,name=Master,sub=Balancer ... Other interesting looking beans are: Class Name: Master,sub=AssignmentManger Object Name: Hadoop:service=DataNode,name=Master,sub=AssignmentManger ... and Object Name: Hadoop:service=DataNode,name=RegionServer,sub=Server Class Name: sun.management.GarbageCollectorImpl The code emitting this stuff seems right using the raw Iterator returned by the bean query. Let me see about how registration is done. > TestStochasticBalancerJmxMetrics.testJmxMetrics_PerTableMode:183 NullPointer > ---------------------------------------------------------------------------- > > Key: HBASE-14423 > URL: https://issues.apache.org/jira/browse/HBASE-14423 > Project: HBase > Issue Type: Sub-task > Components: test > Reporter: stack > Assignee: stack > Fix For: 2.0.0, 1.3.0 > > Attachments: 14423.patch > > > Two of my test runs failed with this. > The mbean is not available. > Let me add dumping of what mbeans are available and then adding in a retry > ... guessing they are slow getting published. Extra printout will give more > clues. -- This message was sent by Atlassian JIRA (v6.3.4#6332)