Sangjin Lee commented on YARN-3334:

Per Zhijie's suggestion, sounds like this code will be removed, so the problem 
will be moot.

Just for completeness, you're right that my suggestion has the other side of 
the issue where a client that is not started can be made available. If we were 
to fix this method properly (which we won't do), I would suggest still 
initializing it before putting it into the map and then if the thread loses the 
race to putIfAbsent() stopping the client (GC will be ok).

> [Event Producers] NM TimelineClient life cycle handling and container metrics 
> posting to new timeline service.
> --------------------------------------------------------------------------------------------------------------
>                 Key: YARN-3334
>                 URL: https://issues.apache.org/jira/browse/YARN-3334
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: nodemanager
>    Affects Versions: YARN-2928
>            Reporter: Junping Du
>            Assignee: Junping Du
>         Attachments: YARN-3334-demo.patch, YARN-3334-v1.patch, 
> YARN-3334-v2.patch, YARN-3334-v3.patch
> After YARN-3039, we have service discovery mechanism to pass app-collector 
> service address among collectors, NMs and RM. In this JIRA, we will handle 
> service address setting for TimelineClients in NodeManager, and put container 
> metrics to the backend storage.

This message was sent by Atlassian JIRA

Reply via email to