hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Amar Kamat (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-4288) java.lang.NullPointerException is observed in Jobtracker log while call heartbeat
Date Mon, 29 Sep 2008 06:34:44 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-4288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12635302#action_12635302
] 

Amar Kamat commented on HADOOP-4288:
------------------------------------

I think something like this *might* have happened
1) User _user_ submitted a job, say _job1_
2) {{CapacityTaskScheduler.TaskSchedulingMgr.jobAdded()}} gets invoked. {{numJobs}} for _user_
is set to 1 and {{numRunningTasks}} is set to 0.
3) Another job was submitted by _user_, say _job2_. Because of a bug in {{CapacityTaskScheduler.TaskSchedulingMgr.jobAdded()}},
{{numJobs}} for _user_ will still be 1.
4) _job1_ completes and {{CapacityTaskScheduler.TaskSchedulingMgr.jobRemoved()}} is invoked.
All entries for _user_ in {{numJobs}} and {{numRunningTasks}} gets removed.
5) Any update for _job2_ will result into NPE as {{numRunningTasks}} is missing an entry for
_user_.

Karam, could you plz verify this using the following steps
1) Find out the job for which the error occured, look at the attempt for which the update
was called.
2) Find out the user of the job, say $user.
3) Check if there were other jobs submitted by the same user and if they have finished
4) If in DEBUG mode, check if you see the following line
{code}
JOb to be removed for user $user
No more jobs for user, number of users
{code}

> java.lang.NullPointerException is observed in Jobtracker log while   call heartbeat
> -----------------------------------------------------------------------------------
>
>                 Key: HADOOP-4288
>                 URL: https://issues.apache.org/jira/browse/HADOOP-4288
>             Project: Hadoop Core
>          Issue Type: Bug
>          Components: contrib/capacity-sched
>    Affects Versions: 0.19.0
>            Reporter: Karam Singh
>            Assignee: Amar Kamat
>
> 2008-09-26 09:33:08,190 INFO org.apache.hadoop.ipc.Server: IPC Server handler 5 on 34441,
call heartbeat(org.apache.hadoop.mapred.TaskTrackerStatus@177a878, false, true, 100) from
<ipaddress:port>: error: java.io.IOException: java.lang.NullPointerException  
> is observed in Jobtracker log.

-- 
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