[ 
https://issues.apache.org/jira/browse/YARN-4287?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14976386#comment-14976386
 ] 

Nathan Roberts commented on YARN-4287:
--------------------------------------

Thanks [~leftnoteasy] for the quick responses. 

{quote}
I think instead of scaling, I suggest to simply cap rack/offswitch delay by the 
cluster size, so:

    rack-delay = min(offswitch, node-locality-delay, cluserSize)
    offswitch-delay = min(offswitch, clusterSize)
    The scaling behavior could be hard to explain to end users.
{quote}

I agree that it's not as easy to describe. BUT, the problem I have is that I 
don't know how to deal with the common case of someone wanting 
node-locality-delay to be based on the size of the cluster. What we do is set 
node-locality-delay to something guaranteed to be larger than the cluster, 
knowing  the scheduler will automatically lower it to the size of the cluster. 
This works great for a single delay on any size cluster. However, it's 
impossible to describe two different delays using this same approach. For 
example, I might always want node-locality-delay to be 10% less than 
rack-locality-delay.  Maybe we should specify rack-locality-delay as a 
percentage above node-locality-delay ( 10%)? Still a little hard to describe 
though. 

> Capacity Scheduler: Rack Locality improvement
> ---------------------------------------------
>
>                 Key: YARN-4287
>                 URL: https://issues.apache.org/jira/browse/YARN-4287
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: capacityscheduler
>    Affects Versions: 2.7.1
>            Reporter: Nathan Roberts
>            Assignee: Nathan Roberts
>         Attachments: YARN-4287-v2.patch, YARN-4287-v3.patch, YARN-4287.patch
>
>
> YARN-4189 does an excellent job describing the issues with the current delay 
> scheduling algorithms within the capacity scheduler. The design proposal also 
> seems like a good direction.
> This jira proposes a simple interim solution to the key issue we've been 
> experiencing on a regular basis:
>  - rackLocal assignments trickle out due to nodeLocalityDelay. This can have 
> significant impact on things like CombineFileInputFormat which targets very 
> specific nodes in its split calculations.
> I'm not sure when YARN-4189 will become reality so I thought a simple interim 
> patch might make sense. The basic idea is simple: 
> 1) Separate delays for rackLocal, and OffSwitch (today there is only 1)
> 2) When we're getting rackLocal assignments, subsequent rackLocal assignments 
> should not be delayed
> Patch will be uploaded shortly. No big deal if the consensus is to go 
> straight to YARN-4189. 



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

Reply via email to