hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joep Rottinghuis (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-6058) Support for listing all applications i.e /apps
Date Fri, 13 Jan 2017 19:43:26 GMT

    [ https://issues.apache.org/jira/browse/YARN-6058?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15822243#comment-15822243
] 

Joep Rottinghuis commented on YARN-6058:
----------------------------------------

Note that we've slowly drifted in terminology and I think in our understanding of what is
proposed by [~vrushalic].
Vrushali suggested to add the existing framework types for each application in a flow-run
to the flow run and flow activity tables. Each and every framework that happened to have an
application in a flow run (and on that specific date for the flow activity table) would be
marked as a special column.

We're not discussing adding a new single flow type for a flow. The types are simply the "enumeration"
of framework types for each app in the flow.

Therefore Tez, MR, Spark and all other framework specific UIs would then pass in their own
framework type to the queries for flow runs and flow activity and retrieve only those flow
runs which contain an application of their type. When they then retrieve applications, an
additional filtering would be needed.

> Support for listing all applications i.e /apps
> ----------------------------------------------
>
>                 Key: YARN-6058
>                 URL: https://issues.apache.org/jira/browse/YARN-6058
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelinereader
>            Reporter: Rohith Sharma K S
>            Assignee: Rohith Sharma K S
>            Priority: Critical
>              Labels: yarn-5355-merge-blocker
>
> Primary use case for /apps is many execution engines runs on top of YARN example, Tez,
MR. These engines will have their own UI's which list specific type of entities which are
published by them Ex: DAG entities. 
> But, these UI's do not aware of either userName or flowName or applicationId which are
submitted by these engines.
> Currently, given that user do not aware of user, flownName, and applicationId, then he
can not retrieve any entities. 
> By supporting /apps with filters, user can list of application with given ApplicationType.
These applications can be used for retrieving engine specific entities like DAG. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
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