Github user sryza commented on a diff in the pull request:
https://github.com/apache/spark/pull/578#discussion_r12036556
--- Diff: core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala ---
@@ -124,6 +124,8 @@ object SparkSubmit {
}
} else if (clusterManager == YARN) {
childMainClass = "org.apache.spark.deploy.yarn.Client"
+ // Special flag to avoid deprecation warnings at the client
+ sysProps("SPARK_SUBMIT_YARN") = "true"
--- End diff --
Even if it's a hack / workaround, I think it's probably better to work with
the assumption that it'll end up sticking around for a while. With that in
mind, a system property that signifies we started something from spark-submit
makes more sense to me than a system property that signifying we started from
spark-submit AND happen to be running YARN. Especially if we'll need a similar
one for standalone mode. Do you think SPARK_SUBMIT_YARN would be less likely
to be abused than SPARK_SUBMIT?
That said, it's a tiny issue, and I don't feel super strongly, so if you
feel differently, go ahead with the current patch.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
