[
https://issues.apache.org/jira/browse/MAPREDUCE-5466?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vinod Kumar Vavilapalli updated MAPREDUCE-5466:
-----------------------------------------------
Status: Open (was: Patch Available)
The analysis looks good to me.
Essentially, we are no longer letting AMs which are erring out anyways to write
history file. The patch needs more changes though
- In case of rebooting AMs, we should skip writing history files except for
the last AM attempt.
- Even jobs failing because of an ERROR event should behave the same. All AMs
except the last retry should skip writing history file.
- In TestJobHistoryEventHandler, make sure that processDoneFilesCalled is true
for successful/failed/killed jobs.
Also, please do some manual testing and report results. It shouldn't be too
difficult to reproduce this with and without the patch.
Thanks!
> Historyserver does not refresh the result of restarted jobs after RM restart
> ----------------------------------------------------------------------------
>
> Key: MAPREDUCE-5466
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5466
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: yeshavora
> Assignee: Jian He
> Attachments: MAPREDUCE-5466.patch
>
>
> Restart RM when sort job is running and verify that the job passes
> successfully after RM restarts.
> Once the job finishes successfully, run job status command for sort job. It
> shows "Job state =FAILED". Job history server does not update the result for
> the job which restarted after RM restart.
> hadoop job -status job_1375923346354_0003
> 13/08/08 01:24:13 INFO mapred.ClientServiceDelegate: Application state is
> completed. FinalApplicationStatus=SUCCEEDED. Redirecting to job history server
> Job: job_1375923346354_0003
> Job File:
> hdfs://host1:port1/history/done/2013/08/08/000000/job_1375923346354_0003_conf.xml
> Job Tracking URL :
> http://historyserver:port2/jobhistory/job/job_1375923346354_0003
> Uber job : false
> Number of maps: 80
> Number of reduces: 1
> map() completion: 0.0
> reduce() completion: 0.0
> Job state: FAILED
> retired: false
> reason for failure: There are no failed tasks for the job. Job is failed due
> to some other reason and reason can be found in the logs.
> Counters not available. Job is retired.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
