[
https://issues.apache.org/jira/browse/YARN-6593?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16095566#comment-16095566
]
Konstantinos Karanasos commented on YARN-6593:
----------------------------------------------
Thanks for the feedback, [~leftnoteasy] and [~chris.douglas]!
I will address all your comments and upload a new version.
bq. I'm not sure when you plan to use
SingleConstraintTransformer/SpecializedConstraintTransformer. If you agree that
they should not be user-facing, do you think we should move them to a separate
JIRA while doing implementation?
I agree they should not be user-facing -- I will fix the annotations.
The transformations will, for example, be used at the scheduler side, if the
scheduler wants to do something with specific cardinality or target
constraints. I would suggest to keep it at this JIRA so that we have all these
related changes clubbed together.
It seems that the problem of updating constraints would require more
discussion. Since this patch is already introducing a lot of new stuff, I would
also suggest to push the update to future work, after we discuss what the exact
semantics of updating constraints are. A first solution to the problem would
be, like you say, to invalidate an existing scheduling request and send a new
one.
> [API] Introduce Placement Constraint object
> -------------------------------------------
>
> Key: YARN-6593
> URL: https://issues.apache.org/jira/browse/YARN-6593
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Konstantinos Karanasos
> Assignee: Konstantinos Karanasos
> Fix For: 3.0.0-alpha3
>
> Attachments: YARN-6593.001.patch, YARN-6593.002.patch,
> YARN-6593.003.patch, YARN-6593.004.patch
>
>
> This JIRA introduces an object for defining placement constraints.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
