Github user skonto commented on a diff in the pull request:
https://github.com/apache/spark/pull/21669#discussion_r199941392
--- Diff:
resource-managers/kubernetes/core/src/main/scala/org/apache/spark/deploy/k8s/Constants.scala
---
@@ -65,11 +65,13 @@ private[spark] object Constants {
val ENV_CLASSPATH = "SPARK_CLASSPATH"
val ENV_DRIVER_BIND_ADDRESS = "SPARK_DRIVER_BIND_ADDRESS"
val ENV_SPARK_CONF_DIR = "SPARK_CONF_DIR"
+ val ENV_SPARK_USER = "SPARK_USER"
--- End diff --
I guess this is for setting the correct user. But I think hadoop libs
should pick the correct user in SparkContext where `Utils.getCurrentUserName()`
is used. In addition I think we should allow the pods to run with any username
(eg. with customized images) and we should have a security context per
container like: `SecurityContextBuilder().withRunAsUser()`. In the latter
scenario hadoop libraries will pick that logged in user as well.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
