You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Varun Saxena (JIRA)" <ji...@apache.org> on 2017/10/21 06:06:40 UTC
[jira] [Updated] (YARN-6888) Refactor AppLevelTimelineCollector
such that RM does not have aggregator threads created
[ https://issues.apache.org/jira/browse/YARN-6888?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Varun Saxena updated YARN-6888:
-------------------------------
Fix Version/s: 2.9.0
> Refactor AppLevelTimelineCollector such that RM does not have aggregator threads created
> ----------------------------------------------------------------------------------------
>
> Key: YARN-6888
> URL: https://issues.apache.org/jira/browse/YARN-6888
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: timelineserver
> Reporter: Vrushali C
> Assignee: Vrushali C
> Fix For: 2.9.0, YARN-5355, 3.0.0-beta1, YARN-5355-branch-2
>
> Attachments: YARN-6888-YARN-5355.0001.patch
>
>
> Currently both RM and NM use the same AppLevelTimelineCollector class. The NM requires aggregator threads per application so that it can perform in memory aggregation for application metrics but the RM does not need this. Since they share the code, RM has a bunch of "TimelineCollector Aggregation" threads created (one per running app).
> Filing jira to refactor AppLevelTimelineCollector such that RM does not have aggregator threads created.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org