[
https://issues.apache.org/jira/browse/SPARK-31142?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17071444#comment-17071444
]
Dongjoon Hyun commented on SPARK-31142:
---------------------------------------
This issue is resolved as `Invalid`. Please see the discussion on the PR.
> Remove useless conf set in pyspark context
> ------------------------------------------
>
> Key: SPARK-31142
> URL: https://issues.apache.org/jira/browse/SPARK-31142
> Project: Spark
> Issue Type: Improvement
> Components: PySpark, Spark Core
> Affects Versions: 3.1.0
> Reporter: Jackey Lee
> Priority: Major
>
> In Pyspark/context.py, we extract the configuration of the
> "spark.executorEnv" prefix from conf and put it in env. But in
> PythonWorkerFactory, we have already obtained these environment variables
> from ProgressBuilder, thus we do not need the extra copy.
> In addition, in some special environment deployments, user-configured
> environment variables such as "spark.executorEnv.Path=/usr/lib:$PATH" will be
> automatically parsed as actual machine confs. If we still directly copy and
> overwrite it, the configuration will be a wrong config.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
