[ 
https://issues.apache.org/jira/browse/YARN-7859?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16348106#comment-16348106
 ] 

wangwj commented on YARN-7859:
------------------------------

    Hi [~yufeigu],Thank you for your reply.
    I think that no violation of the core part of fair scheduler if enable this 
feature(deadline).Because I only Improve the phenomenon that one queue has not 
shceduled for a long time.When the resource of a queue is sufficient,and this 
queue  has not shceduled for a long time.We mandatory scheduler this queue 
once.I do not modify the fairShare of any schedulable.The fairness of each 
schedulable is still in accordance with the previous algorithm.
    I think it is necessary to explain the phenomenon of scheduling 
starvation.The apps in one queue are pending and the resource used of this 
queue is smaller than minResource of the queue.
    Thanks...
    
   

> New feature: add queue scheduling deadLine in fairScheduler.
> ------------------------------------------------------------
>
>                 Key: YARN-7859
>                 URL: https://issues.apache.org/jira/browse/YARN-7859
>             Project: Hadoop YARN
>          Issue Type: New Feature
>          Components: fairscheduler
>    Affects Versions: 3.0.0
>            Reporter: wangwj
>            Priority: Major
>              Labels: fairscheduler, features, patch
>             Fix For: 3.0.0
>
>         Attachments: YARN-7859-v1.patch, log, screenshot-1.png, 
> screenshot-3.png
>
>   Original Estimate: 24h
>  Remaining Estimate: 24h
>
>      As everyone knows.In FairScheduler the phenomenon of queue scheduling 
> starvation often occurs when the number of cluster jobs is large.The App in 
> one or more queue are pending.So I have thought a way to solve this 
> problem.Add queue scheduling deadLine in fairScheduler.When a queue is not 
> scheduled for FairScheduler within a specified time.We mandatory scheduler it!
>     Now the way of community solves queue scheduling to starvation is preempt 
> container.But this way may increases the failure rate of the job.
>     On the basis of the above, I propose this issue...



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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

Reply via email to