[
https://issues.apache.org/jira/browse/YARN-5932?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sunil G updated YARN-5932:
--------------------------
Attachment: YARN-5932.v0.patch
Updating a version 0 patch to capture the flow.
Changes done:
- Currently this patch handles changes only for Capacity Scheduler
- A new pre-validate check is added for moveApplication in CapacityScheduler
and Leaf/ParentQueue. This helps to ensure that the target queue can accept
this application or not. Any failure could be handled at this stage itself.
- StateStore also updated with the queue name to which this app is moved to. We
could update submission context for same.
- Finally perform real move operation. Ideally there should not be any
exception thrown out. However with some corner cases, there is very less
chance to happen this. In such cases, we can revert from StateStore.
cc/[~jianhe][~rohithsharma]
> Retrospect moveToQueue in align with YARN-5611
> ----------------------------------------------
>
> Key: YARN-5932
> URL: https://issues.apache.org/jira/browse/YARN-5932
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacity scheduler, resourcemanager
> Reporter: Sunil G
> Assignee: Sunil G
> Attachments: YARN-5932.v0.patch
>
>
> All dynamic api's of an application's state change could follow a general
> design approach. Currently priority and app timeouts are following this
> approach all corner cases.
> *Steps*
> - Do a pre-validate check to ensure that changes are fine.
> - Update this information to state-store
> - Perform real move operation and update in-memory data structures.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
