You are viewing a plain text version of this content. The canonical link for it is here.

Posted to yarn-issues@hadoop.apache.org by "Wei Shao (JIRA)" <ji...@apache.org> on 2015/06/21 01:48:01 UTC

[jira] [Updated] (YARN-3808) Proposal of Time Extended Fair Scheduling for YARN

     [ https://issues.apache.org/jira/browse/YARN-3808?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Wei Shao updated YARN-3808:
---------------------------
    Summary: Proposal of Time Extended Fair Scheduling for YARN  (was: Proposal of Time Based Fair Scheduling for YARN)

> Proposal of Time Extended Fair Scheduling for YARN
> --------------------------------------------------
>
>                 Key: YARN-3808
>                 URL: https://issues.apache.org/jira/browse/YARN-3808
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: fairscheduler, scheduler
>            Reporter: Wei Shao
>         Attachments: ProposalOfTimeBasedFairSchedulingForYARN-V1.02.pdf
>
>
> This proposal talks about the issues of YARN fair scheduling policy, and tries to solve them by YARN-3806 and the new scheduling policy called time based fair scheduling.
> Time based fair scheduling policy is proposed to enforces time based fairness among users. For example, if two users share the cluster weekly, each user’s fair share is half of the cluster per week. At a particular week, if the first user has used the whole cluster for first half of the week, then in second half of the week, second user will always have priority to use cluster resources since the first user has used up its fair share of the cluster already.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)