Arun Suresh updated YARN-5637:
    Attachment: YARN-5637.006.patch

Simplifying patch based on some offline discussions with [~jianhe]

Essentially, the behavior now is:
If the new process dies after launch,
* If _autoCommit is true_, it will simply follow the rules specified by the 
*ContainerRetryContext* found in the new launch context (terminate immediately, 
restart, restart infinitely etc.)
* If _autoCommit is false_, It will do exactly as above but at the end of the 
retry sequence, instead of terminating, it will ReInitialize back to the old 
launch context.

If the new process startsup fine:
* If _autoCommit is false_, the AM has the option of either
** Committing the container if it feels the re-initialization is successful. 
The rollback context is deleted.
** Explicitly Rollback the container to previous launch context if it feels 
there is something wrong
** Not do anything. The rollback context will hang around until another re-init 
is called.

> Changes in NodeManager to support Container rollback and 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-5637.004.patch, YARN-5637.005.patch, 
> YARN-5637.006.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

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