[
https://issues.apache.org/jira/browse/HADOOP-5352?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12677377#action_12677377
]
Vinod K V commented on HADOOP-5352:
-----------------------------------
In one case, TT the localization took a long time because of DistributedCache
localizing files from DFS, and just before task timeout interval elapsed, the
task was put to running state. TT in its offerService loop detected this task
as running, compared the TIP's lastProgressReport to now and killed it as
unresponsive. The TIP's lastProgressReport was set way back when the TIP itself
got created in the TIP'sconstructor and the task didn't get to report it's
progress yet.
> Setting lastProgressReport time in TIP's constructor causes TT to wrongly
> kill tasks.
> -------------------------------------------------------------------------------------
>
> Key: HADOOP-5352
> URL: https://issues.apache.org/jira/browse/HADOOP-5352
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.20.0
> Reporter: Vinod K V
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
