spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chris Martin (JIRA)" <>
Subject [jira] [Commented] (SPARK-6305) Add support for log4j 2.x to Spark
Date Wed, 22 Aug 2018 20:09:00 GMT


Chris Martin commented on SPARK-6305:

I don't think that's such a big deal so long as Spark can have a hard dependency on log4j
directly (as opposed to being slf4j with log4j as the default backend).  At the moment I'm
assuming the former and it's relatively easy (although not as easy as log4j 1.x) to adjust
the log level as needed.  If one needs to consider other logging backends (e.g. JUL) then
all this gets more complicated as the log level manipulation needs to work for each backend
we support  and if we need to consider log4j 1.x as yet another backend then my approach
is not going to work!

For waht ti's worth this is one of the reasons why I'm a bit confused by logging.scala. 
There's some code (and a comment) in there which implies that users should be able to swap
out for JUL if they want.  I can't see how this will ever work through as all the code for
switching log levels is assuming log4j loggers!

> Add support for log4j 2.x to Spark
> ----------------------------------
>                 Key: SPARK-6305
>                 URL:
>             Project: Spark
>          Issue Type: Improvement
>          Components: Build
>            Reporter: Tal Sliwowicz
>            Priority: Minor
> log4j 2 requires replacing the slf4j binding and adding the log4j jars in the classpath.
Since there are shaded jars, it must be done during the build.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message