hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Konstantinos Karanasos (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-6594) [API] Introduce SchedulingRequest object
Date Wed, 02 Aug 2017 19:23:01 GMT

    [ https://issues.apache.org/jira/browse/YARN-6594?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16111595#comment-16111595

Konstantinos Karanasos commented on YARN-6594:

Thanks for the explanation, [~jianhe].

I see why key-value pairs will be needed to select containers. BTW when would an AM want to
select containers with specific attributes though? Like when you are a service and want to
update specific containers only?

Like you say, we could change the container tags to be key-value pairs. I was trying to think
of use cases that this would be really needed and I didn't find many. For example, we could
have {{memory_critical=high}} instead of just {{memory-critical}}. But then the constraint
API would get too complicated, because we would need to teach people that they can have constraints
where only the key matters (as in "I want allocations that are memory_critical, no matter
the value") and others that the value matters too.

Adding they key-values in YARN-6593 is not hard, but I would suggest to keep it as is for
the first version. 
For this JIRA, we can make the tags to be key-values, but only if we think it is useful.
Another problem, like you say, is that the current tags refer to allocations, while it seems
that your key-value attributes would refer to containers.

Let me know what you guys think.

> [API] Introduce SchedulingRequest object
> ----------------------------------------
>                 Key: YARN-6594
>                 URL: https://issues.apache.org/jira/browse/YARN-6594
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Konstantinos Karanasos
>            Assignee: Konstantinos Karanasos
>         Attachments: YARN-6594.001.patch
> This JIRA introduces a new SchedulingRequest object.
> It will be part of the {{AllocateRequest}} and will be used to define sizing (e.g., number
of allocations, size of allocations) and placement constraints for allocations.
> Applications can use either this new object (when rich placement constraints are required)
or the existing {{ResourceRequest}} object.

This message was sent by Atlassian JIRA

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

View raw message