hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sangjin Lee (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-3030) set up ATS writer with basic request serving structure and lifecycle
Date Tue, 20 Jan 2015 21:00:35 GMT

    [ https://issues.apache.org/jira/browse/YARN-3030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14284396#comment-14284396

Sangjin Lee commented on YARN-3030:

Thanks for the comments [~zjshen]. Some response below.

bq. I can see BaseAggregatorService is the abstraction for the latter piece. However, for
the former piece we don't have such kind of abstraction, but the implementation embedded in
the per-node aggregator.

I believe we're basically on the same page. Actually the intent is that BaseAggregatorService
*is* the place for the abstraction for the former. It's just that that piece is left unimplemented
mostly due to the fact that the object model has not been finalized/implemented (see BaseAggregatorService#postEntities
and BaseAggregatorService#postEntitiesAsync). The idea is that BaseAggregatorService should
really include all the common and generic logic of the timeline aggregator, and the app-level
aggregator and the per-node aggregator on top (and also the future RM aggregator) would be
pretty thin extensions.

The reason why the per-node aggregator is the REST endpoint and not BaseAggregatorService
is because there may be multiple BaseAggregatorService instances that can be instantiated
dynamically inside a per-node aggregator. For simplicity, the entry point (per-node aggregator)
is the REST endpoint, but its job is to simply route the calls to the right base aggregator
service and manage their lifecycle as they come and go.

Hope these answer your feedback. Let me know.

bq. For per-node aggregator, we make want to implement {{AuxiliaryService}} interface, such
that it can be installed to NM as the auxiliary service.

Sounds good. Let me work on that. So it would be instead of extending the {{CompositeService}},

> set up ATS writer with basic request serving structure and lifecycle
> --------------------------------------------------------------------
>                 Key: YARN-3030
>                 URL: https://issues.apache.org/jira/browse/YARN-3030
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelineserver
>            Reporter: Sangjin Lee
>            Assignee: Sangjin Lee
>         Attachments: YARN-3030.001.patch
> Per design in YARN-2928, create an ATS writer as a service, and implement the basic service
structure including the lifecycle management.
> Also, as part of this JIRA, we should come up with the ATS client API for sending requests
to this ATS writer.

This message was sent by Atlassian JIRA

View raw message