[jira] [Commented] (MAPREDUCE-3921) MR AM should act on the nodes liveliness information when nodes go up/down/unhealthy

["Bikas Saha (Commented) (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/MAPREDUCE-3921?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13226794#comment-13226794 ] 

Bikas Saha commented on MAPREDUCE-3921:
---------------------------------------

Attaching a patch that builds on MAPREDUCE-3353
1) RMContainerAllocator receives node updates along with allocated containers
2) It sends KILL event to map task attempts running on unusable nodes
3) It sends a JobUpdatedNode event to JobImpl
4) JobImpl maintains a mapping of nodes to successful task attempts that have run on them
5) On receiving updated nodes JobImpl sends KILL event to map task attempts from 4)
6) Successful task completions retro actively get to the KILLED state if their successful task attempt is the same as the task attempt in 5). They reschedule another attempt.
                
> MR AM should act on the nodes liveliness information when nodes go up/down/unhealthy
> ------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-3921
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-3921
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: mr-am, mrv2
>    Affects Versions: 0.23.0
>            Reporter: Vinod Kumar Vavilapalli
>            Assignee: Bikas Saha
>             Fix For: 0.23.2
>
>         Attachments: MAPREDUCE-3921-branch-0.23.patch
>
>


--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira