[
https://issues.apache.org/jira/browse/YARN-5561?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15472930#comment-15472930
]
Varun Saxena edited comment on YARN-5561 at 9/8/16 6:14 AM:
------------------------------------------------------------
bq. So the new addition looks fine to me.
Looks fine to me as well.
bq. Do we want to reorganize the code in a way consistent with this list?
Well the current organization is based on what we are retrieving. That is, all
endpoints for fetching entities are together, for fetching apps are together
and so on.
We can follow approach suggested by you as well. I do not have a strong opinion
on either. So I will leave it as it is. Lets see what others think.
bq. discuss the endpoints that marked as weird
These endpoints were added to get all apps belonging to a flow so we skip the
flow run section. There were use cases to fetch all apps within a flow in case
run id is not known. Refer to [~vrushalic]'s [comment on YARN-3864 |
https://issues.apache.org/jira/browse/YARN-3864?focusedCommentId=14940489&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14940489]
We also plan to list all apps for a user or queue in future as well.
And based on use case of Rohith maybe list all apps within a cluster as well.
However in my personal opinion that may not be necessary. You can check with
New Web UI folks.
was (Author: varun_saxena):
bq. So the new addition looks fine to me.
Looks fine to me as well.
bq. Do we want to reorganize the code in a way consistent with this list?
Well the current organization is based on what we are retrieving that is all
endpoints for fetching entities are together, for fetching apps are together
and so on.
We can follow approach suggested by you as well. I do not have a strong opinion
on either. So I will leave it as it is. Lets see what others think.
bq. discuss the endpoints that marked as weird
These endpoints were added to get all apps belonging to a flow so we skip the
flow run section. There were use cases to fetch all apps within a flow in case
run id is not known. Refer to [~vrushalic]'s [comment on YARN-3864 |
https://issues.apache.org/jira/browse/YARN-3864?focusedCommentId=14940489&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14940489]
We also plan to list all apps for a user or queue in future as well.
And based on use case of Rohith maybe list all apps within a cluster as well.
However in my personal opinion that may not be necessary. You can check with
New Web UI folks.
> [Atsv2] : Support for ability to retrieve apps/app-attempt/containers and
> entities via REST
> -------------------------------------------------------------------------------------------
>
> Key: YARN-5561
> URL: https://issues.apache.org/jira/browse/YARN-5561
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: timelinereader
> Reporter: Rohith Sharma K S
> Assignee: Rohith Sharma K S
> Attachments: YARN-5561.patch, YARN-5561.v0.patch
>
>
> ATSv2 model lacks retrieval of {{list-of-all-apps}},
> {{list-of-all-app-attempts}} and {{list-of-all-containers-per-attempt}} via
> REST API's. And also it is required to know about all the entities in an
> applications.
> It is pretty much highly required these URLs for Web UI.
> New REST URL would be
> # GET {{/ws/v2/timeline/apps}}
> # GET {{/ws/v2/timeline/apps/\{app-id\}/appattempts}}.
> # GET
> {{/ws/v2/timeline/apps/\{app-id\}/appattempts/\{attempt-id\}/containers}}
> # GET {{/ws/v2/timeline/apps/\{app id\}/entities}} should display list of
> entities that can be queried.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
