Added back the initial spark job before starting streaming receivers

streaming/src/main/scala/spark/streaming/NetworkInputTracker.scala

@@ -141,7 +141,7 @@ class NetworkInputTracker(
       }
       // Run the dummy Spark job to ensure that all slaves have registered.
       // This avoids all the receivers to be scheduled on the same node.
-      //ssc.sparkContext.makeRDD(1 to 100, 100).map(x => (x, 1)).reduceByKey(_ + _, 20).collect()
+      ssc.sparkContext.makeRDD(1 to 50, 50).map(x => (x, 1)).reduceByKey(_ + _, 20).collect()
 
       // Distribute the receivers and start them
       ssc.sparkContext.runJob(tempRDD, startReceiver)