Sangjin Lee commented on YARN-3039:

[~zjshen], [~djp], regarding the idea about having IPC from NM to the per-app 
collector, I don't think that will work with a special container use case. The 
special container for the per-app collector will bind to a port for RPC that 
will not be determined until the time the collector binds to it. So it's 
basically a chicken-and-egg problem: NM doesn't know the RPC port for the 
per-all collector in the special container until ... the special containers 
tells it. This is not a problem with the current per-node collector container 

Although it's a little roundabout, I don't see a fundamental problem with 
having the per-app collector (or the collection of them) sending its location 
to the NM once it's up. It's actually conceptually simpler, and it should work 
in all 3 modes (aux service, standalone per-node daemon, and special container).

> [Aggregator wireup] Implement ATS app-appgregator service discovery
> -------------------------------------------------------------------
>                 Key: YARN-3039
>                 URL: https://issues.apache.org/jira/browse/YARN-3039
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelineserver
>            Reporter: Sangjin Lee
>            Assignee: Junping Du
>         Attachments: Service Binding for applicationaggregator of ATS 
> (draft).pdf, Service Discovery For Application Aggregator of ATS (v2).pdf, 
> YARN-3039-no-test.patch, YARN-3039-v2-incomplete.patch, 
> YARN-3039-v3-core-changes-only.patch, YARN-3039-v4.patch, YARN-3039-v5.patch
> Per design in YARN-2928, implement ATS writer service discovery. This is 
> essential for off-node clients to send writes to the right ATS writer. This 
> should also handle the case of AM failures.

This message was sent by Atlassian JIRA

Reply via email to