Arun Suresh commented on YARN-2885:

Thanks for the review [~leftnoteasy],

bq. Do you have real use case that distributed scheduler needs to set different 
properties such as DIST_SCHEDULING_MIN_MEMORY..
We had included this so that we can centrally (via the RM) control the max/min 
capability of a container that is allocatable via Distributed Scheduling.. 
which could be different from the Scheduler min/max.

bq. First constructor of ApplicationMasterService, should use name ..
Good catch.. will fix that

bq. You can add a isDistributedSchedulingEnabled method to YarnConfiguration..
Actually, I had thought about that.. but if you notice, we are actually 
querying the configuration object only once in the serviceinit of the 
{{NodeManager}} and the value is passed on to all the other required classes 
via constructor etc.. So I did not find much value in adding that the 

If you are ok with the rest of the patch and agree with the approach... I shall 
clean it up a bit more with proper javadocs etc.

> Create AMRMProxy request interceptor for distributed scheduling decisions for 
> queueable containers
> --------------------------------------------------------------------------------------------------
>                 Key: YARN-2885
>                 URL: https://issues.apache.org/jira/browse/YARN-2885
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: nodemanager, resourcemanager
>            Reporter: Konstantinos Karanasos
>            Assignee: Arun Suresh
>         Attachments: YARN-2885-yarn-2877.001.patch, 
> YARN-2885-yarn-2877.002.patch, YARN-2885-yarn-2877.full-2.patch, 
> YARN-2885-yarn-2877.full-3.patch, YARN-2885-yarn-2877.full.patch, 
> YARN-2885-yarn-2877.v4.patch, YARN-2885_api_changes.patch
> We propose to add a Local ResourceManager (LocalRM) to the NM in order to 
> support distributed scheduling decisions. 
> Architecturally we leverage the RMProxy, introduced in YARN-2884. 
> The LocalRM makes distributed decisions for queuable containers requests. 
> Guaranteed-start requests are still handled by the central RM.

This message was sent by Atlassian JIRA

Reply via email to