[ 
https://issues.apache.org/jira/browse/YARN-5637?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Arun Suresh updated YARN-5637:
------------------------------
    Attachment: YARN-5637.003.patch

Updating patch.. thanks again for the review [~jianhe]..

bq. Here, we could make reInitEvent.getResourceSet() be merged with existing 
resourceSet.localizedResource upfront, so that both oldResourceSet and 
newResourceSet contain full copy of resources, rather than delta.
This was actually intentional. Consider the case where the original process has 
many resources to Localize but the upgrade launch script just needs a binary 
change in addition to the existing resources. If the resourceSets were merged 
upfront, then in the _ReInitializeContainerTransition_, the 
_ContainerLocalizationRequestEvent_ that gets sent would include ALL the 
resources, instead of just the single resource. The Container will have to 
remain in the *REINITIALIZING* state till it receives _RESOURCE_LOCALIZED_ 
events for all the resources in the combined resultset before being able to 
launch.

bq. the container.reInitContext!= null check is not needed..
Think we do, else it might cause an NPE when the _LaunchTransition_ happens as 
part of the initial container startup.

bq. I found the resourceSet is also not updated when rollback in 
RetryFailureTransition
Good catch... I also like your refactoring.. i've incorporated it in the latest 
patch

> Changes in NodeManager to support Container upgrade and rollback/commit
> -----------------------------------------------------------------------
>
>                 Key: YARN-5637
>                 URL: https://issues.apache.org/jira/browse/YARN-5637
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Arun Suresh
>            Assignee: Arun Suresh
>         Attachments: YARN-5637.001.patch, YARN-5637.002.patch, 
> YARN-5637.003.patch
>
>
> YARN-5620 added support for re-initialization of Containers using a new 
> launch Context.
> This JIRA proposes to use the above feature to support upgrade and subsequent 
> rollback or commit of the upgrade.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to