[
https://issues.apache.org/jira/browse/HADOOP-5374?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Amareshwari Sriramadasu updated HADOOP-5374:
--------------------------------------------
Status: Open (was: Patch Available)
cancelling patch, current code accepts FAILED_UNCLEAN/KILLED_UNCLEAN states and
launches cleanup attempt, though previous state is FAILED. This should not
happen, because FAILED->FAILED_UNCLEAN->FAILED updates will fail the task
twice, thus doubling the counters and punishing TT twice.
I would say, It should not accept any state if the previous state is
FAILED/KILLED.
> NPE in JobTracker.getTasksToSave() method
> -----------------------------------------
>
> Key: HADOOP-5374
> URL: https://issues.apache.org/jira/browse/HADOOP-5374
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.19.0
> Reporter: Vinod K V
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.19.2
>
> Attachments: patch-5374-1.txt
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
