Github user pwendell commented on the pull request:
https://github.com/apache/spark/pull/148#issuecomment-37769373
It's unfortunate that Hadoop publishes the log4j.properties file directly
in it's jars. This is exactly why we've avoided doing this in Spark because it
creates a weird situation where you can't easily control logging preferences.
@sryza @mridulm - I think what we need is a precedence order here:
- If a user explicitly gives a log4j.properties, use that (@mridul - I
didn't actually know we allowed this in the YARN launcher, where is the code
relevant to that?)
- If a user jar has a log4j.properties file in it, we should use that.
- If not, we can use the YARN cluster defaults when running on YARN.
---
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.
---
