hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Subru Krishnan (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-6871) Add a new REST API GetAppReport call lighter and faster than the current one
Date Tue, 25 Jul 2017 22:54:00 GMT

    [ https://issues.apache.org/jira/browse/YARN-6871?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16100910#comment-16100910
] 

Subru Krishnan commented on YARN-6871:
--------------------------------------

[~giovanni.fumarola], +1 on the proposal. Can we use the generic *deSelects* param introduced
in YARN-6280 to achieve the same?

[~cltlfcjin]/[~sunilg], thoughts?

> Add a new REST API GetAppReport call lighter and faster than the current one
> ----------------------------------------------------------------------------
>
>                 Key: YARN-6871
>                 URL: https://issues.apache.org/jira/browse/YARN-6871
>             Project: Hadoop YARN
>          Issue Type: New Feature
>          Components: resourcemanager, router
>            Reporter: Giovanni Matteo Fumarola
>
> This jira tracks the effort to create a new REST API similar to the current GetAppReport
but lighter and faster.
> With the current one we are facing a scalability issues.
> E.g. with ~500 applications running the AppReport can reach up to 300MB in size due to
the {{ResourceRequest}} in the {{AppInfo}}.
> The new method returns for each application the following and essential information:
> * The application id;
> * The application name;
> * The application state according to the ResourceManager - valid values are members of
the YarnApplicationState enum: NEW, NEW_SAVING, SUBMITTED, ACCEPTED, RUNNING, FINISHED, FAILED,
KILLED;
> * The final status of the application if finished - reported by the application itself
- valid values are: UNDEFINED, SUCCEEDED, FAILED, KILLED;
> * The web URL that can be used to track the application;
> * Detailed diagnostics information;
> * The URL of the application master container logs;
> * The nodes http address of the application master;
> * The progress of the application as a percent.
> Yarn RM will return the new result faster and it will use less compute cycles to create
the report and it will improve the YARN RM and Client's performances.



--
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


Mime
View raw message