Github user tdas commented on a diff in the pull request:
https://github.com/apache/spark/pull/7276#discussion_r34752244
--- Diff:
streaming/src/main/scala/org/apache/spark/streaming/scheduler/ReceiverSchedulingPolicy.scala
---
@@ -0,0 +1,110 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.streaming.scheduler
+
+import scala.collection.mutable
+import scala.util.Random
+
+import org.apache.spark.streaming.scheduler.ReceiverState._
+
+private[streaming] case class ReceiverTrackingInfo(
+ receiverId: Int,
+ state: ReceiverState,
+ scheduledLocations: Option[Seq[String]],
+ runningLocation: Option[String])
+
+private[streaming] trait ReceiverSchedulingPolicy {
+
+ /**
+ * Return a list of candidate executors to run the receiver. If the list
is empty, the caller can
+ * run this receiver in arbitrary executor.
+ */
+ def scheduleReceiver(
+ receiverId: Int,
+ preferredLocation: Option[String],
+ receiverTrackingInfoMap: Map[Int, ReceiverTrackingInfo],
+ executors: Seq[String]): Seq[String]
+}
+
+/**
+ * A ReceiverScheduler trying to balance executors' load. Here is the
approach to schedule executors
+ * for a receiver.
+ * <ol>
+ * <li>
+ * If preferredLocation is set, preferredLocation should be one of the
candidate executors.
+ * </li>
+ * <li>
+ * Every executor will be assigned to a weight according to the
receivers running or scheduling
+ * on it.
+ * <ul>
+ * <li>
+ * If a receiver is running on an executor, it contributes 1.0 to
the executor's weight.
+ * </li>
+ * <li>
+ * If a receiver is scheduled to an executor but has not yet run,
it contributes
+ * `1.0 / #candidate_executors_of_this_receiver` to the executor's
weight.</li>
+ * </ul>
+ * At last, if there are more than 3 idle executors (weight = 0),
returns all idle executors.
+ * Otherwise, we only return 3 best options according to the weights.
+ * </li>
+ * </ol>
+ *
+ */
+private[streaming] class LoadBalanceReceiverSchedulingPolicyImpl extends
ReceiverSchedulingPolicy {
--- End diff --
LoadBalanceReceiverSchedulingPolicyImpl -->
LoadBalance*d*ReceiverSchedulingPolicy
Why add impl? Either use `ReceiverSchedulingPolicyImpl` or
`LoadBalancedReceiverSchedulingPolicy` I am fine either ways. In face, the
first one might be better since there is only one implementation
---
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.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
