[
https://issues.apache.org/jira/browse/SPARK-5388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14306899#comment-14306899
]
Dale Richardson commented on SPARK-5388:
----------------------------------------
Heh Andrew, definitely starting to look a bit more Rest-like in the protocol!
Http Delete should be used for your kill request - it is considered best
practice
The primary resource you are dealing with is a submission - this should form
the base of your url structure.
For a rest protocol, actions/verbs are used to affect these resources - so they
are mapped to to the HTTP operations of GET/POST/DELETE/HEAD/OPTIONS etc,
against the resources defined by the full url.
Full URLs serve to identify the resources that these actions are performed on.
GET/DELETE are used where the full identity of the resource is known at the
time of generating the request, POST is used when you may not know the address
of the resource at the time of generating the request (eg When submitting a
program to run, you will not know submission id because it is returned by the
request)
So, taking this into account:
RequestSubmitDriver → POST /submission/create
RequestKillDriver → DELETE /submission/[submissionId]
RequestDriverStatus → GET /submission/[submissionId]/status - The resource is
the submission, so the current status of the submission in a sub-resource of
the submission, other sub entries such as
/submission/[submissionId]/performanceCounters
could be added in the future without affecting existing clients.
> Provide a stable application submission gateway in standalone cluster mode
> --------------------------------------------------------------------------
>
> Key: SPARK-5388
> URL: https://issues.apache.org/jira/browse/SPARK-5388
> Project: Spark
> Issue Type: Bug
> Components: Spark Core
> Affects Versions: 1.2.0
> Reporter: Andrew Or
> Assignee: Andrew Or
> Priority: Blocker
> Attachments: stable-spark-submit-in-standalone-mode-2-4-15.pdf
>
>
> The existing submission gateway in standalone mode is not compatible across
> Spark versions. If you have a newer version of Spark submitting to an older
> version of the standalone Master, it is currently not guaranteed to work. The
> goal is to provide a stable REST interface to replace this channel.
> For more detail, please see the most recent design doc attached.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
