Github user vanzin commented on a diff in the pull request: https://github.com/apache/spark/pull/22897#discussion_r230491208 --- Diff: resource-managers/kubernetes/docker/src/main/dockerfiles/spark/entrypoint.sh --- @@ -96,22 +96,6 @@ case "$SPARK_K8S_CMD" in "$@" ) ;; - driver-py) --- End diff -- > if a user upgraded their spark-submit version to 3.0 but didn't upgrade the version of Spark in their docker image I don't think Spark really supports that in any case. It is expected that you're using the same version of Spark in the spark-submit side and on the other side, regardless of the cluster manager. e.g., imagine client mode in the scenario you're talking about. But definitely, the contract here needs to be well defined (I've asked for this a long time ago and that's one of the reasons why k8s is still experimental).
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org