Github user vanzin commented on the pull request:
https://github.com/apache/spark/pull/3409#issuecomment-65144899
@zhzhan sorry I should have expanded a little bit on my questions.
What I'm trying to get at is two things:
* if mapred-site.xml supports referencing system properties, the same most
probably applies to hdfs-site.xml and core-site.xml, right? And those are read
by the executors.
* it feels weird to add an AM-specific option that would be used in two
different contexts. In client mode, you'd add these extra arguments to the AM.
In cluster mode, you'd add these plus `spark.driver.extraJavaOptions`. And if
the bullet above is correct, you'd need to duplicate the same properties in the
executor-related config option.
So I wonder if there shouldn't be a single option that could apply to all
three? Also, also assuming the first bullet above is correct, then this ceases
to be exclusive to Yarn, and both standalone and mesos could benefit from the
same feature.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
