gengliangwang commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2099078293
@dongjoon-hyun Thanks, I will add a configuration about the unification.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
dongjoon-hyun commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2098985001
WDYT, @gengliangwang ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific
dongjoon-hyun commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2098984770
I understand your point, @mridulm . Do you use the following syntax?
```
spark.sparkContext.setLocalProperty("mdc." + name, "value")
```
Initially, I thought it's
mridulm commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2097607873
So if I understood right, this as a backwardly incompatible change done for
consistency of naming for Structured logging ?
If yes, IMO it is preferable to revert this and not break
gengliangwang commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2097446465
@mridulm As the task name MDC is frequently showing in the logs; I would say
this is necessary for the new logging framework. After the renaming, the MDC
names are consistent and
mridulm commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2097430754
Is this change strictly necessary would be the question ... if it is, we can
evaluate it in that context.
If not and is a nice to have, it is better not to make breaking changes
which
gengliangwang commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2097287397
> That change was not visible to end users, as there was not release made -
right ?
Yes you are right.
Are you ok with the changes in this PR? If not, please let me know
mridulm commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2097182888
That change was not visible to end users, as there was not release made -
right ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on
gengliangwang commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2096535872
> Wouldn't this not impact/break existing log config files where users are
customizing the template ?
Yes, and this is documented on `docs/configuration.md`. If necessary, I
mridulm commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2094831806
Wouldn't this not impact/break existing log config files where users are
customizing the template ?
--
This is an automated message from the Apache Git Service.
To respond to the
dongjoon-hyun commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2094521040
Merged to master for Apache Spark 4.0.0-preview.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL
dongjoon-hyun closed pull request #46386: [SPARK-48131][Core] Unify MDC key
`mdc.taskName` and `task_name`
URL: https://github.com/apache/spark/pull/46386
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to
gengliangwang commented on PR #46386:
URL: https://github.com/apache/spark/pull/46386#issuecomment-2094434379
I suggest having this merged before the Spark 4.0 preview release.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
gengliangwang opened a new pull request, #46386:
URL: https://github.com/apache/spark/pull/46386
### What changes were proposed in this pull request?
Currently there are two MDC keys for task name:
* `mdc.taskName`, which is introduced in
14 matches
Mail list logo