[
https://issues.apache.org/jira/browse/BEAM-2591?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16357726#comment-16357726
]
Ben Sidhom commented on BEAM-2591:
----------------------------------
[~kenn]
Is this still the case? Do we want the Python SDK to spin up a Flink deployment
and then submit jobs to it? If not, then we need the endpoint of an
already-running Job API anyway. This is no different for Flink than any other
runner.
The main distinction here is if we want the FlinkRunner to run locally, in
which case it might be reasonable to have Flink create an in-process
"deployment" and execute against it.
> Python shim for submitting to FlinkRunner
> -----------------------------------------
>
> Key: BEAM-2591
> URL: https://issues.apache.org/jira/browse/BEAM-2591
> Project: Beam
> Issue Type: Sub-task
> Components: runner-flink, sdk-py-core
> Reporter: Kenneth Knowles
> Priority: Major
> Labels: portability
>
> Whatever the result of https://s.apache.org/beam-job-api, Python users will
> need to be able to pass --runner=FlinkRunner and have it work.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
