hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Panagiotis Garefalakis (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-7653) Rack cardinality support for AllocationTagsManager
Date Tue, 02 Jan 2018 14:12:00 GMT

    [ https://issues.apache.org/jira/browse/YARN-7653?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16308092#comment-16308092

Panagiotis Garefalakis commented on YARN-7653:

Hello [~leftnoteasy],  I agree with the title change.
Regarding the node group support: 
* in the discussion above we agreed that we need to at least support Rack as it already defined
in our API
* in the committed patch the CountedTags inner class is generic with the goal to support any
arbitrary node group. The only thing we would add is an extra data structure keeping a group
to CountedTags mapping (in that scenario RACK would be just a specific node group)
* to keep things simple since we dont have arbitrary groups so far this extra mapping is not
there - as we would also need a way to define/add/remove node groups -  but I would be happy
to work on that if we want to support it

> Rack cardinality support for AllocationTagsManager
> --------------------------------------------------
>                 Key: YARN-7653
>                 URL: https://issues.apache.org/jira/browse/YARN-7653
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Panagiotis Garefalakis
>            Assignee: Panagiotis Garefalakis
>             Fix For: YARN-6592
>         Attachments: YARN-7653-YARN-6592.001.patch, YARN-7653-YARN-6592.002.patch, YARN-7653-YARN-6592.003.patch
> AllocationTagsManager currently supports node and cluster-wide tag cardinality retrieval.
> If we want to support arbitrary node-groups/scopes for our placement constraints TagsManager
should be extended to provide such functionality.
> As a first step we need to support RACK scope cardinality retrieval (as defined in our
> i.e. how many "spark" containers are currently running on "RACK-1"

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