You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@beam.apache.org by "Kenneth Knowles (JIRA)" <ji...@apache.org> on 2019/03/19 03:34:00 UTC

[jira] [Commented] (BEAM-2593) Python shim for submitting to SparkRunner

    [ https://issues.apache.org/jira/browse/BEAM-2593?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16795625#comment-16795625 ] 

Kenneth Knowles commented on BEAM-2593:
---------------------------------------

[~angoenka] I think we are not doing these shims, right? Are we trying to let Python users write things like {{--runner=SparkRunner}}? My understanding is that instead we write {{--runner=PortableRunner}} and specify an endpoint. Is there value still in writing helpers that launch a local job proxy for the intended runner?

> Python shim for submitting to SparkRunner
> -----------------------------------------
>
>                 Key: BEAM-2593
>                 URL: https://issues.apache.org/jira/browse/BEAM-2593
>             Project: Beam
>          Issue Type: Sub-task
>          Components: runner-spark, sdk-py-core
>            Reporter: Kenneth Knowles
>            Priority: Major
>              Labels: portability, triaged
>
> Whatever the result of https://s.apache.org/beam-job-api, Python users will need to be able to pass --runner=SparkRunner and have it work.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)