[
https://issues.apache.org/jira/browse/MAPREDUCE-4252?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13444383#comment-13444383
]
Bikas Saha commented on MAPREDUCE-4252:
---------------------------------------
We could move the check for TaskType after the check for successful task
attempt has been made. This is similar to the pattern for
MapRetroActiveFailureTransition(). So the fix is a simple downward movement of
the check for TaskType.
The check for whether the task is MAP should still be there (and is also there
for MapRetroActiveFailureTransition()) because after a REDUCE has completed
with SUCCEEDED there is no reason to fail it because its outputs are already
preserved in HDFS. So any attempt to fail it would be a bug in the code.
We could also potentially improve the check of a successful attempt in
MapRetroActiveKilledTransition() by using the condition in
MapRetroActiveFailureTransition().
I opened YARN-61 to track this. Currently assigned to me because I wrote
MapRetroactiveKilledTransition. [~chaku88] Feel free to assign this to yourself.
> MR2 job never completes with 1 pending task
> -------------------------------------------
>
> Key: MAPREDUCE-4252
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-4252
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.1
> Reporter: Tom White
> Assignee: Tom White
> Fix For: 0.23.3, 2.1.0-alpha, 3.0.0
>
> Attachments: MAPREDUCE-4252.patch, MAPREDUCE-4252.patch,
> MAPREDUCE-4252.patch, MAPREDUCE-4252.patch, MapReduce.png
>
>
> This was found by ATM:
> bq. I ran a teragen with 1000 map tasks. Many task attempts failed, but after
> 999 of the tasks had completed, the job is now sitting forever with 1 task
> "pending".
--
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
