hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Amar Kamat (JIRA)" <j...@apache.org>
Subject [jira] Updated: (MAPREDUCE-1246) JobTrackerStatistics/JobTrackerInstrumentation's tracker count will be incorrect if a live tracker is restarted on the same port
Date Fri, 27 Nov 2009 10:51:39 GMT

     [ https://issues.apache.org/jira/browse/MAPREDUCE-1246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Amar Kamat updated MAPREDUCE-1246:
----------------------------------

    Description: Consider a case where an active tracker is restarted and the tracker restarts
on the same port. In such a case the jobtracker we will invoke JobTracker.loseTaskTracker()
from processHeartbeat(). Later on the tracker is added back (addNewTracker()) which will update
the tracker count in metrics andinstrumentation. But the problem is that when the old tracker
is lost the metrics and instrumentation's tracker counts are not decremented resulting into
incorrect trackers count.   (was: Consider a case where an active tracker is restarted and
the tracker restarts on the same port. In such a case the jobtracker we will invoke JobTracker.loseTracker()
from processHeartbeat(). Later on the tracker is added back (addNewTracker()) which will update
the metrics and instrumentation. But the problem is that when the old tracker is lost the
metric/instrumentation counts are not accounted for. )
        Summary: JobTrackerStatistics/JobTrackerInstrumentation's tracker count will be incorrect
if a live tracker is restarted on the same port  (was: JobTrackerStatistics/JobTrackerInstrumentation
counts will be incorrect if a live tracker is restarted on the same port)

> JobTrackerStatistics/JobTrackerInstrumentation's tracker count will be incorrect if a
live tracker is restarted on the same port
> --------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-1246
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1246
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: jobtracker
>            Reporter: Amar Kamat
>
> Consider a case where an active tracker is restarted and the tracker restarts on the
same port. In such a case the jobtracker we will invoke JobTracker.loseTaskTracker() from
processHeartbeat(). Later on the tracker is added back (addNewTracker()) which will update
the tracker count in metrics andinstrumentation. But the problem is that when the old tracker
is lost the metrics and instrumentation's tracker counts are not decremented resulting into
incorrect trackers count. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message