[ https://issues.apache.org/jira/browse/HADOOP-10090?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13824029#comment-13824029 ]
Chris Nauroth commented on HADOOP-10090: ---------------------------------------- Thanks, Ivan. This looks right to me. Thanks for catching the locking order in {{updateJmxCache}} too. (I had forgotten that this would be trying to take the lock on {{this}} before calling {{getMetrics}}, which would have invalidated the logic I described above.) I verified that the new test is passing on Mac. Just one very minor nit: {{TestMetricsSourceAdapter#testJmx}} could say {{throws Exception}} instead of listing the various exception types. Since it's a test, and any exception is unexpected, I think throwing the base {{Exception}} class is acceptable, and it shrinks the code size a little. Shall we do a trunk patch too? Like Luke said, this seems to be a moot point due to YARN-1043, but perhaps it's best to keep the code lines in sync as much as possible. > Jobtracker metrics not updated properly after execution of a mapreduce job > -------------------------------------------------------------------------- > > Key: HADOOP-10090 > URL: https://issues.apache.org/jira/browse/HADOOP-10090 > Project: Hadoop Common > Issue Type: Bug > Components: metrics > Affects Versions: 1.2.1 > Reporter: Ivan Mitic > Assignee: Ivan Mitic > Attachments: HADOOP-10090.branch-1.2.patch, > HADOOP-10090.branch-1.patch, OneBoxRepro.png > > > After executing a wordcount mapreduce sample job, jobtracker metrics are not > updated properly. Often times the response from the jobtracker has higher > number of job_completed than job_submitted (for example 8 jobs completed and > 7 jobs submitted). > Issue reported by Toma Paunovic. -- This message was sent by Atlassian JIRA (v6.1#6144)