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 "Vrushali C (JIRA)" <ji...@apache.org> on 2018/03/08 22:33:00 UTC
[jira] [Commented] (YARN-6936) [Atsv2] Retrospect storing entities
into sub application table from client perspective
[ https://issues.apache.org/jira/browse/YARN-6936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16392026#comment-16392026 ]
Vrushali C commented on YARN-6936:
----------------------------------
Currently, ATSv2 stores information in an additional table (sub application table) if the doAs user and AM user are different. This I believe is one of the Tez usage situations, DAGs run as individual users via the same AM (different AM user).
Since DAGs or queries are not a primary YARN concept, they are not treated as first class citizens by timeline service. Timeline Service v2 will treat all entities as being inside of an application scope. Reading entities outside of application scope may not be a desirable recommendation.
Let's chat with Tez community to talk about the querying of entities done by Tez so that we can ascertain that the ATSv2 APIs address it and/or arrive at a generic recommendation for all frameworks that use ATsv2.
> [Atsv2] Retrospect storing entities into sub application table from client perspective
> --------------------------------------------------------------------------------------
>
> Key: YARN-6936
> URL: https://issues.apache.org/jira/browse/YARN-6936
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Rohith Sharma K S
> Priority: Major
>
> Currently YARN-6734 stores entities into sub application table only if doAs user and submitted users are different. This holds good for Tez kind of use cases. But AM runs as same as submitted user like MR also need to store entities in sub application table so that it could read entities without application id.
> This would be a point of concern later stages when ATSv2 is deployed into production. This JIRA is to retrospect decision of storing entities into sub application table based on client side configuration driven rather than user driven.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org