Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/186#discussion_r11713563
--- Diff: core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala
---
@@ -1152,13 +1006,158 @@ class DAGScheduler(
}
def stop() {
- if (eventProcessActor != null) {
- eventProcessActor ! StopDAGScheduler
- }
+ logInfo("Stopping DAGScheduler")
+ dagSchedulerActorSupervisor ! PoisonPill
taskScheduler.stop()
}
}
+private[scheduler] class DAGSchedulerActorSupervisor(dagScheduler:
DAGScheduler)
+ extends Actor with Logging {
+
+ override val supervisorStrategy =
+ OneForOneStrategy() {
+ case x: Exception => {
+ logError("eventProcesserActor failed due to the error %s; shutting
down SparkContext"
+ .format(x.getMessage))
+ dagScheduler.doCancelAllJobs()
+ dagScheduler.sc.stop()
+ Stop
+ }
+ }
+
+ def receive = {
+ case p: Props => sender ! context.actorOf(p)
+ case _ =>
+ }
+
+ dagScheduler.eventProcessActor = context.actorOf(
+ Props(new DAGSchedulerEventProcessActor(dagScheduler)))
--- End diff --
I think this should be moved to L77. That way `eventProcessActor` doesn't
need to be a var. (It was given a default value of `_` originally because we
wanted to set it only in `start()`, but now that's gone we can just set it on
construction of DAGScheduler)
---
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.
---
