You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Geoffrey Mon (JIRA)" <ji...@apache.org> on 2017/01/26 16:32:24 UTC
[jira] [Commented] (FLINK-5183) [py] Support multiple jobs per
Python plan file
[ https://issues.apache.org/jira/browse/FLINK-5183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15839961#comment-15839961 ]
Geoffrey Mon commented on FLINK-5183:
-------------------------------------
I'm currently working on this at https://github.com/geofbot/flink/tree/FLINK-5183. So far I've managed to get it to work except for a few issues:
* Execution environments need to be able to get a unique ID that can be used to identify it between Java and Python. At the moment, these IDs are manually assigned.
* PythonPlanBinder does not exit when Python process exits, but instead waits indefinitely for more jobs
* Global variables used to run operators in Python for skipping over execution environments that do not contain the operator to be run (there should be a better solution)
> [py] Support multiple jobs per Python plan file
> -----------------------------------------------
>
> Key: FLINK-5183
> URL: https://issues.apache.org/jira/browse/FLINK-5183
> Project: Flink
> Issue Type: Improvement
> Components: Python API
> Affects Versions: 1.1.3
> Reporter: Geoffrey Mon
> Priority: Minor
>
> Support running multiple jobs per Python plan file.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)