Yikun commented on a change in pull request #35422:
URL: https://github.com/apache/spark/pull/35422#discussion_r805128127
##########
File path: resource-managers/kubernetes/core/pom.xml
##########
@@ -29,8 +29,25 @@
<name>Spark Project Kubernetes</name>
<properties>
<sbt.project.name>kubernetes</sbt.project.name>
+ <volcano.exclude>**/*Volcano*.scala</volcano.exclude>
</properties>
+ <profiles>
+ <profile>
+ <id>volcano</id>
Review comment:
@HyukjinKwon Thanks for raising the question, let me enumerate the
difference between `built-in module` and `profile` first:
1. Built-in (https://github.com/apache/spark/pull/35495):
- `volcano-model-v1beta1` deps for spark-kubernetes module
- `volcano-client` deps for spark-integrations-test module
2. Profile (https://github.com/apache/spark/pull/35422):
- `volcano-model-v1beta1` deps for spark-kubernetes module with volcano
profile
- `volcano-client` deps for spark-integrations-test module with volcano
profile
- Add exclude info in `sbt` AND `maven`.
- Add `-Pvolcano` in everywhere.
- Required seprate file for volcano (including test/code) to help add
profile filter in mvn/sbt, we couldn't put volcano test mix in existing test
unless `extends`.
I have to admit profile module is really a little bit complex for me even I
have made efforts to complete the separating work.
So if it possible, I personally suggest that have it as bulit-in module and
still required to use the separate class / file for Volcano to reduce the
impact. Of course, I'd like to hear some input from @dongjoon-hyun at this
time. WDYT?
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
