Github user mridulm commented on a diff in the pull request:
https://github.com/apache/spark/pull/8760#discussion_r46248240
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/BlacklistTracker.scala ---
@@ -0,0 +1,253 @@
+/*
+ * 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.concurrent.TimeUnit
+
+import scala.collection.mutable
+
+import org.apache.spark.SparkConf
+import org.apache.spark.Success
+import org.apache.spark.TaskEndReason
+import org.apache.spark.annotation.DeveloperApi
+import org.apache.spark.util.SystemClock
+import org.apache.spark.util.ThreadUtils
+import org.apache.spark.util.Utils
+import org.apache.spark.util.Clock
+
+/**
+ * BlacklistTracker is design to track problematic executors and node on
application level.
+ * It is shared by all TaskSet, so that once a new TaskSet coming, it
could be benefit from
+ * previous experience of other TaskSet.
+ *
+ * Once task finished, the callback method in TaskSetManager should update
+ * executorIdToFailureStatus Map.
+ */
+private[spark] class BlacklistTracker(
+ sparkConf: SparkConf,
+ clock: Clock = new SystemClock()) extends BlacklistCache{
+ // maintain a ExecutorId --> FailureStatus HashMap
+ private val executorIdToFailureStatus: mutable.HashMap[String,
FailureStatus] = mutable.HashMap()
+
+ // Apply Strategy pattern here to change different blacklist detection
logic
+ private val strategy = BlacklistStrategy(sparkConf)
+
+ // A daemon thread to expire blacklist executor periodically
+ private val scheduler =
ThreadUtils.newDaemonSingleThreadScheduledExecutor(
+ "spark-scheduler-blacklist-expire-timer")
+
+ private val recoverPeriod = sparkConf.getTimeAsSeconds(
+ "spark.scheduler.blacklist.recoverPeriod", "60s")
+
+ def start(): Unit = {
+ val scheduleTask = new Runnable() {
+ override def run(): Unit = {
+ Utils.logUncaughtExceptions(expireExecutorsInBlackList())
+ }
+ }
+ scheduler.scheduleAtFixedRate(scheduleTask, 0L, recoverPeriod,
TimeUnit.SECONDS)
+ }
+
+ def stop(): Unit = {
+ scheduler.shutdown()
+ scheduler.awaitTermination(10, TimeUnit.SECONDS)
+ }
+
+ // The actual implementation is delegated to strategy
+ /** VisibleForTesting */
+ private[scheduler] def expireExecutorsInBlackList(): Unit = synchronized
{
+ val updated =
strategy.expireExecutorsInBlackList(executorIdToFailureStatus, clock)
+ if (updated) {
+ invalidateCache()
+ }
+ }
--- End diff --
@kayousterhout Would be good to have your thoughts here since I am fairly
out of touch with scheduler now.
If I read it correctly, multiple methods in this class are directly called
from scheduler as part of identifying, starting, updating tasks to be run from
TaskSetManager/scheduler and they all use the same lock as
updateFailedExecutors - which can actually run fairly non trivial computation
as number of failures increase in the system. Is this a concern or is it
alleviated somehow and I missed it ?
---
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
