[ 
https://issues.apache.org/jira/browse/YARN-6594?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16112097#comment-16112097
 ] 

Jian He commented on YARN-6594:
-------------------------------

bq. BTW when would an AM want to select containers with specific attributes 
though? Like when you are a service and want to update specific containers only?
Yep, e.g. I want to upgrade containers with specific label annotated. Also, 
since I want to annotate metaInfo (most importantly, container-name) to the 
container, if it's just a bag of strings, I won't know which string is for 
container-name etc. unless I build my own parsing at client side. 

One other option is to have underlying proto implementation as a map, but user 
facing java API keeps as set and does transformation underneath. In case 
there's a such use-case for it to be a map, we can extend the user-facing java 
API, but the proto can keep the same. FYI, this 
[link|https://kubernetes.io/docs/concepts/configuration/assign-pod-node/#inter-pod-affinity-and-anti-affinity-beta-feature]
 shows some use-case about how kubernetes models as key/value pairs for 
affinity/anti-affinity between containers. 
In any case, I'm ok to keep as-is because the current API is also sufficient 
for my use-case. 

It just occurred to me that, the design doc didn't cover changing/add the 
container allocation tags, was that discussed ? I think at least we need to 
account for such requirement when implementing. I do think this as a valid 
use-case in future. 


 

> [API] Introduce SchedulingRequest object
> ----------------------------------------
>
>                 Key: YARN-6594
>                 URL: https://issues.apache.org/jira/browse/YARN-6594
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Konstantinos Karanasos
>            Assignee: Konstantinos Karanasos
>         Attachments: YARN-6594.001.patch
>
>
> This JIRA introduces a new SchedulingRequest object.
> It will be part of the {{AllocateRequest}} and will be used to define sizing 
> (e.g., number of allocations, size of allocations) and placement constraints 
> for allocations.
> Applications can use either this new object (when rich placement constraints 
> are required) or the existing {{ResourceRequest}} object.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to