[
https://issues.apache.org/jira/browse/SPARK-2849?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andrew Or updated SPARK-2849:
-----------------------------
Description: We currently ignore all spark.driver.* for client mode. I
understand that these are originally intended for the driver when it is
launched on one of the worker machines (i.e. for cluster mode). However, from
the names spark.driver.* themselves, the user should expect these options to
work regardless of what deploy mode they're running on. This has been a source
of confusion on the mailing list. (was: We currently ignore all \-\-driver-*
options in client mode. Meanwhile, elsewhere we tell the user to use
--driver-class-path if SPARK_CLASSPATH is set, for instance.
More generally, I understand that these are originally intended for the driver
when it is launched on one of the worker machines (i.e. for cluster mode).
However, from the names \-\-driver-* themselves, the user should expect these
options to work regardless of what deploy mode they're running on.)
> bin/spark-submit should respect spark.driver.* for client mode
> --------------------------------------------------------------
>
> Key: SPARK-2849
> URL: https://issues.apache.org/jira/browse/SPARK-2849
> Project: Spark
> Issue Type: Bug
> Components: Spark Core
> Affects Versions: 1.0.2
> Reporter: Andrew Or
> Priority: Blocker
> Fix For: 1.1.0
>
>
> We currently ignore all spark.driver.* for client mode. I understand that
> these are originally intended for the driver when it is launched on one of
> the worker machines (i.e. for cluster mode). However, from the names
> spark.driver.* themselves, the user should expect these options to work
> regardless of what deploy mode they're running on. This has been a source of
> confusion on the mailing list.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
