[
https://issues.apache.org/jira/browse/YARN-1011?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Karthik Kambatla updated YARN-1011:
-----------------------------------
Attachment: yarn-1011-design-v2.pdf
Based on the comments here and learnings from the paper Bikas pointed out, I
have updated the design with more details. Most of the new details are in the
section named Scheduling (Section 3.1 in page 2).
[~bikassaha], [~nroberts], [~leftnoteasy], [~elgoiri] - could you give it a
quick read through, and see if it is missing any major things. I am fairly
certain we ll learn more as we implement/use the feature.
> [Umbrella] Schedule containers based on utilization of currently allocated
> containers
> -------------------------------------------------------------------------------------
>
> Key: YARN-1011
> URL: https://issues.apache.org/jira/browse/YARN-1011
> Project: Hadoop YARN
> Issue Type: New Feature
> Reporter: Arun C Murthy
> Attachments: yarn-1011-design-v0.pdf, yarn-1011-design-v1.pdf,
> yarn-1011-design-v2.pdf
>
>
> Currently RM allocates containers and assumes resources allocated are
> utilized.
> RM can, and should, get to a point where it measures utilization of allocated
> containers and, if appropriate, allocate more (speculative?) containers.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
