[ 
https://issues.apache.org/jira/browse/YARN-6059?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15868890#comment-15868890
 ] 

Konstantinos Karanasos commented on YARN-6059:
----------------------------------------------

Just checked the updated patch, [~hrsharma].
Some comments:
* The paused container has to go past the {{scheduleContainer()}} method of 
{{ContainerScheduler}} to reach your newly-added codepath. For this to happen, 
resources have to be available for the container to be scheduled, whereas what 
we really want to do is simply kill the recovered paused container. You see 
what I mean?
* {{RecoverPausedContainerLaunch}}: by sending the ContainerExitEvent inside 
the try statement, in case there is a problem in the commands before, the 
ContainerExitEvent will never be sent. So let's move it after the try/catch 
statement.
* {{ContainerLaunch}}: 
** storeContainerQueued -> storeContainerPaused (LOG.warn message under it 
needs fixing too);
** I think you need to add a similar call in the resumeContainer() to put it 
back to the running state.
* {{NMLeveldbStateStoreService}}: I don't see why we need the if's in lines 
248, 252-254, and 258-260. I think the rcs.status can be nothing but REQUESTED 
at this point.
* Fix checkstyle issues.
* Can you run the failing tests locally without your changes and make sure they 
are also failing there as well?

Nits:
* {{ContainerImpl}}: let's not introduce the new line (it can cause merge 
conflicts on other people)
* {{ContainersLauncher}}: please put the RECOVER_PAUSED_CONTAINER case below 
the RECOVER_CONTAINER

> Update paused container state in the state store
> ------------------------------------------------
>
>                 Key: YARN-6059
>                 URL: https://issues.apache.org/jira/browse/YARN-6059
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Hitesh Sharma
>            Assignee: Hitesh Sharma
>         Attachments: YARN-5216-YARN-6059.001.patch, 
> YARN-6059-YARN-5972.001.patch, YARN-6059-YARN-5972.002.patch, 
> YARN-6059-YARN-5972.003.patch, YARN-6059-YARN-5972.004.patch
>
>




--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org

Reply via email to