You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by GitBox <gi...@apache.org> on 2021/01/12 16:36:25 UTC
[GitHub] [spark] tgravescs commented on pull request #31151: [SPARK-34088][CORE] Rename all decommission configurations to use the same namespace "spark.decommission.*"
tgravescs commented on pull request #31151:
URL: https://github.com/apache/spark/pull/31151#issuecomment-758779625
Right, so this is going against the current name standards as was brought up in that discussion. the current format of these configs matches with that format where spark.executor, spark.workers, etc are the prefixes. The documentation in many cases categorizes them into sections by these prefixes. I get that naming them by feature may make it easier in some cases, but I also see where naming them as it makes it easier in others. I know that when I setup worker configuration I look for all spark.worker configs and those might go into separate file just for workers, whereas I put spark.executor configs in a file to be loaded by all applications, or when I am looking at things that might be used to optimize shuffle, I look at all spark.shuffle configs. My point is one person may think its easier this new way, others might think its easier as is. We have been using spark.executor, spark.worker, etc for a long time so changing that will introduce some confusion, especially if only
done for a single feature.
So I'm against make this change unless we decide in general to change our naming rules for configs and think it should be decided on by the devs and documented appropriately. These configs were put in with the feature and reviewed by committers we should not be changing them last minute without good cause.
----------------------------------------------------------------
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 comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org