Github user pavel-sakun closed the pull request at:
https://github.com/apache/spark/pull/5202
---
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
Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-87403389
Mind closing this PR?
---
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
GitHub user pavel-sakun opened a pull request:
https://github.com/apache/spark/pull/5202
SPARK-6549 - Spark console logger logs to stderr by default
Changed default log4j configuration to log into stdout by default
You can merge this pull request into a Git repository by running:
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86403254
Can one of the admins verify this patch?
---
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
Github user WangTaoTheTonic commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86416244
LGTM
---
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
Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86482979
No, I'm pretty certain this is on purpose. This change makes log output mix
with program output. You can override it in your own logging config if you
like.
---
If
Github user pavel-sakun commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86491813
Isn't it why loggers/levels/categories and all that stuff exist? Stderr as
it's name implies should be used for error messages. In case of submitting job
to yarn
Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86495944
`stderr` is just a name; I read it as a side-channel for output that is
different from program output. It can be for error messages, but is not
inherently only for error
Github user pwendell commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86810509
Changing the default logging behavior like this would break compatibility
for people who rely on the out-of-the-box behavior.
IIRC this change was proposed
Github user WangTaoTheTonic commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86626534
I understand Sean says `This change makes log output mix with program
output.`
Perhaps log the spark message to System.err could not solve the mixed
problem
Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86674350
The question is, what's a reasonable _default_? I don't think the default
can involve logging to files, if that's what you mean. Yes, this is why you
would customize the
Github user vanzin commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86685660
The SparkSubmit stream is on purpose because pyspark expects the very first
line of stdout to be the port number to connect to the py4j gateway. So you
can't print
Github user pavel-sakun commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86525129
Name usually tells something about the purpose of the object, so this
channel was most likely intended for writing error messages so they are not
mess with normal
Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/5202#issuecomment-86544848
Sure, but the name is a POSIX thing from 30 years ago.
I understand that a program with no output might reasonably log to `stdout`
but I don't think it's crazy
14 matches
Mail list logo
