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

Posted to issues@flink.apache.org by "Robert Metzger (JIRA)" <ji...@apache.org> on 2019/02/28 14:27:00 UTC

[jira] [Updated] (FLINK-5409) Separate Job API from the Runtime

     [ https://issues.apache.org/jira/browse/FLINK-5409?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Robert Metzger updated FLINK-5409:
----------------------------------
    Component/s:     (was: Java API)
                 API / DataStream

> Separate Job API from the Runtime
> ---------------------------------
>
>                 Key: FLINK-5409
>                 URL: https://issues.apache.org/jira/browse/FLINK-5409
>             Project: Flink
>          Issue Type: Improvement
>          Components: API / DataStream
>            Reporter: Matt Zimmer
>            Priority: Major
>
> Currently, all of the Flink runtime is visible to jobs.  It will make classpath management easier if only the minimum needed to create processing jobs is loaded in the job ClassLoader.  This would ideally be limited to the job API and jars placed in a folder designated for sharing across jobs (as /lib is now, but it contains the flink-dist_*.jar) and would exclude flink runtime support classes and jars.
> I've discussed this with [~till.rohrmann] and [~rmetzger].



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