Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/2746#discussion_r18739007
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/ExecutorScalingManager.scala ---
@@ -0,0 +1,324 @@
+/*
+ * 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.scheduler
+
+import java.util.{Timer, TimerTask}
+
+import scala.collection.mutable
+
+import org.apache.spark.{Logging, SparkException}
+import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend
+
+/**
+ * An agent that dynamically scales the number of executors based on the
workload.
+ *
+ * The add policy depends on the number of pending tasks. If the queue of
pending tasks has not
+ * been drained for N seconds, then new executors are added. If the queue
persists for another M
+ * seconds, then more executors are added and so on. The number added in
each round increases
+ * exponentially from the previous round until an upper bound on the
number of executors has
+ * been reached.
+ *
+ * The rationale for the exponential increase is twofold: (1) Executors
should be added slowly
+ * in the beginning in case the number of extra executors needed turns out
to be small. Otherwise,
+ * we may add more executors than we need just to remove them later. (2)
Executors should be added
+ * quickly over time in case the maximum number of executors is very high.
Otherwise, it will take
+ * a long time to ramp up under heavy workloads.
+ *
+ * The remove policy is simpler: If an executor has been idle, meaning it
has not been scheduled
+ * to run any tasks, for K seconds, then it is removed. This requires
starting a timer on each
+ * executor instead of just starting a global one as in the add case.
+ *
+ * The relevant Spark properties include the following:
+ * spark.dynamicAllocation.enabled - Whether this feature is enabled
+ * spark.dynamicAllocation.minExecutors - Lower bound on the number of
executors
+ * spark.dynamicAllocation.maxExecutors - Upper bound on the number of
executors
+ * spark.dynamicAllocation.addExecutorThreshold - How long before new
executors are added (N)
+ * spark.dynamicAllocation.addExecutorInterval - How often to add new
executors (M)
+ * spark.dynamicAllocation.removeExecutorThreshold - How long before an
executor is removed (K)
+ *
+ * Synchronization: Because the schedulers in Spark are single-threaded,
contention only arises
+ * if the application itself runs multiple jobs concurrently. Under normal
circumstances, however,
+ * synchronizing each method on this class should not be expensive
assuming biased locking is
+ * enabled in the JVM (on by default for Java 6+). Tighter locks are also
used where possible.
+ *
+ * Note: This is part of a larger implementation (SPARK-3174) and
currently does not actually
+ * request to add or remove executors. The mechanism to actually do this
will be added separately,
+ * e.g. in SPARK-3822 for Yarn.
+ */
+private[scheduler] class ExecutorScalingManager(scheduler:
TaskSchedulerImpl) extends Logging {
--- End diff --
Though I think we need some notion of `executor` in there.
`DynamicExecutorAllocationManager`? `ExecutorAllocationManager`?
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
