Zhijie Shen commented on YARN-3210:

bq. . Yes, it's currently an auxiliary service to the NM, but it can easily be 
started up as a standalone daemon.

I think the idea behind it is that all the node-level aggregator logic should 
be inside TimelineAggregatorsCollection. In aux service mode, 
TimelineAggregatorsCollection is started in 
PerNodeTimelineAggregatorAuxService, while in stand-alone mode, 
TimelineAggregatorsCollection is started as a separate process.

The problem seems to be that the way to start a logic app-level aggregator is 
not decoupled with the aux service. To make it common no matter the aggregator 
is in aux service, standalone process or container, starting the app-level 
aggregator could be treated as a IPC call in Aggregator<->NM communication 

> [Source organization] Refactor timeline aggregator according to new code 
> organization
> -------------------------------------------------------------------------------------
>                 Key: YARN-3210
>                 URL: https://issues.apache.org/jira/browse/YARN-3210
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelineserver
>            Reporter: Li Lu
>            Assignee: Li Lu
>              Labels: refactor
>             Fix For: YARN-2928
>         Attachments: YARN-3210-022715.patch, YARN-3210-030215.patch, 
> YARN-3210-030215_1.patch, YARN-3210-030215_2.patch
> We may want to refactor the code of timeline aggregator according to the 
> discussion of YARN-3166, the code organization for timeline service v2. We 
> need to refactor the code after we reach an agreement on the aggregator part 
> of YARN-3166. 

This message was sent by Atlassian JIRA

Reply via email to