Github user vanzin commented on a diff in the pull request:
https://github.com/apache/spark/pull/1218#discussion_r15546179
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala
---
@@ -63,6 +67,8 @@ private[spark] class SparkDeploySchedulerBackend(
client = new AppClient(sc.env.actorSystem, masters, appDesc, this,
conf)
client.start()
+
+ waitForRegistration()
--- End diff --
See
https://github.com/vanzin/spark/commit/0afd696d2458498cb7db8f1b6f9a18a81c7686f1.
In standalone mode, the call to start() will return before the connection
to the master has been established. This just waits until the master replies
with the application id. This is actually very similar to the code that exists
in the mesos backend.
---
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.
---
