[
https://issues.apache.org/jira/browse/SPARK-6305?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15401129#comment-15401129
]
Sean Owen commented on SPARK-6305:
----------------------------------
Yeah, I tried making Spark use log4j 2 and it was almost intractable. Most of
the problem is maven exclusions to get every single instance of log4j 1.x in
dependencies to map to sl4fj, because there are so many and keep popping up,
and is a little harder because it's one version of log4j remapped to another
via slf4j.
Updating the actual use of log4j 1.x code was hard but not impossible because
there are no direct counterparts to some existing methods. The final issue
wasn't really a Spark issue but a usability one: lots of people's logging
configs would stop working. I abandoned it as a result.
I wonder if it would be simpler to map slf4j to the JDK logger if we need to.
It doesn't solve the remapping problem but at least means we don't put log4j 2
in the way as well.
> Add support for log4j 2.x to Spark
> ----------------------------------
>
> Key: SPARK-6305
> URL: https://issues.apache.org/jira/browse/SPARK-6305
> 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
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
