hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rohith Sharma K S (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-8834) Provide Java client for fetching entities from TimelineReader
Date Wed, 03 Oct 2018 06:19:00 GMT

    [ https://issues.apache.org/jira/browse/YARN-8834?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16636505#comment-16636505

Rohith Sharma K S commented on YARN-8834:

Thanks for the patch.. Few comments
# ClusterId is super important for any REST queries. So take client configuration cluster-id
in serviceInit and use that as default. Else we may need to provide option to provide cluster-id
in API. In short, instead of {{"apps/" + appId.toString()}}, this can be {{"cluster-id/apps/"
+ appId.toString();}}
# I think retrieving specific fields should be controlled by API params. So,   params.add("fields",
"ALL"); could be taken as param? Or default fields could be INFO because retrieving ALL would
consume more memory especially in list of entities. IIRC, for constructing XXXReport, only
INFO field is sufficient.
# For constructing URL, could we use static JOINER? 
# getApplicationEntity has LOG with info method. Could it be removed? 

Leaving apart all my above comments, I am thinking should reader client more abstract to entities
only? I mean just to have getEntity() and getEntities() with much more options to provide
query params. Thoughts? 

> Provide Java client for fetching entities from TimelineReader
> -------------------------------------------------------------
>                 Key: YARN-8834
>                 URL: https://issues.apache.org/jira/browse/YARN-8834
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelinereader
>            Reporter: Rohith Sharma K S
>            Assignee: Abhishek Modi
>            Priority: Critical
>         Attachments: YARN-8834.001.patch, YARN-8834.002.patch
> While reviewing YARN-8303, we felt that it is necessary to provide TimelineReaderClient
which wraps all the REST calls in it so that user can just provide EntityType and EntityId
along with filters.Currently fetching entities from TimelineReader is only via REST call or
somebody need to write java client get entities.
> It is good to provide TimelineReaderClient which fetch entities from TimelineReaderServer.
This will be more useful.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org

View raw message