gaborgsomogyi commented on a change in pull request #25709: [SPARK-29007][STREAMING][MLLIB][TESTS] Enforce not leaking SparkContext in tests which creates new StreamingContext with new SparkContext URL: https://github.com/apache/spark/pull/25709#discussion_r322235691
########## File path: streaming/src/test/scala/org/apache/spark/streaming/LocalStreamingContext.scala ########## @@ -0,0 +1,91 @@ +/* + * 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 + +import org.scalatest.{BeforeAndAfterAll, BeforeAndAfterEach, Suite} + +import org.apache.spark.SparkContext +import org.apache.spark.internal.Logging + +/** + * Manages a local `ssc` `StreamingContext` variable, correctly stopping it after each test. + * Note that it also stops active SparkContext if `stopSparkContext` is set to true (default). + * In most cases you may want to leave it, to isolate environment for SparkContext in each test. + */ +trait LocalStreamingContext extends BeforeAndAfterEach with BeforeAndAfterAll { self: Suite => + + @transient var ssc: StreamingContext = _ + @transient var stopSparkContext: Boolean = true + + override def beforeAll() { + super.beforeAll() + } + + override def afterEach() { + try { + resetStreamingContext() + } finally { + super.afterEach() + } + } + + def resetStreamingContext(): Unit = { + LocalStreamingContext.stop(ssc, stopSparkContext) + ssc = null + } +} + +object LocalStreamingContext extends Logging { + def stop(ssc: StreamingContext, stopSparkContext: Boolean): Unit = { + if (stopSparkContext) { + ensureNoActiveSparkContext(ssc) + } else if (ssc != null) { + ssc.stop(stopSparkContext = false) + } + } + + /** + * Clean up active SparkContext: try to stop first if there's an active SparkContext. + * If it fails to stop, log warning message and clear active SparkContext to avoid + * interfere between tests. + */ + def ensureNoActiveSparkContext(): Unit = { + // if SparkContext is still active, try to clean up + SparkContext.getActive match { + case Some(sc) => + try { + sc.stop() + } catch { + case e: Throwable => + logWarning("Exception trying to stop SparkContext, clear active SparkContext...", e) + SparkContext.clearActiveContext() + } + case _ => + } + } + + def ensureNoActiveSparkContext(ssc: StreamingContext): Unit = { Review comment: This function does a little bit more + I have the feeling this can be merged with `stop`. ``` def stop(ssc: StreamingContext, stopSparkContext: Boolean): Unit = { if (ssc != null) { ssc.stop(stopSparkContext) } ensureNoActiveSparkContext() } ``` BTW, why are these functions are belonging to the object? There are transient vars for `ssc` and `stopSparkContext`. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
