在同一JVM中检测到多个SparkContext

Question

according my last question I have to define the Multiple SparkContext for my unique JVM.

I did it in the next way (using Java):

SparkConf conf = new SparkConf();
conf.setAppName("Spark MultipleContest Test");
conf.set("spark.driver.allowMultipleContexts", "true");
conf.setMaster("local");

After that I create the next source code:

SparkContext sc = new SparkContext(conf);
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

and later in the code:

JavaSparkContext ctx = new JavaSparkContext(conf);
JavaRDD<Row> testRDD = ctx.parallelize(AllList);

After the code executing I got next error message:

16/01/19 15:21:08 WARN SparkContext: Multiple running SparkContexts detected in the same JVM!
org.apache.spark.SparkException: Only one SparkContext may be running in this JVM (see SPARK-2243). To ignore this error, set spark.driver.allowMultipleContexts = true. The currently running SparkContext was created at:
org.apache.spark.SparkContext.<init>(SparkContext.scala:81)
test.MLlib.BinarryClassification.main(BinaryClassification.java:41)
    at org.apache.spark.SparkContext$$anonfun$assertNoOtherContextIsRunning$1.apply(SparkContext.scala:2083)
    at org.apache.spark.SparkContext$$anonfun$assertNoOtherContextIsRunning$1.apply(SparkContext.scala:2065)
    at scala.Option.foreach(Option.scala:236)
    at org.apache.spark.SparkContext$.assertNoOtherContextIsRunning(SparkContext.scala:2065)
    at org.apache.spark.SparkContext$.setActiveContext(SparkContext.scala:2151)
    at org.apache.spark.SparkContext.<init>(SparkContext.scala:2023)
    at org.apache.spark.api.java.JavaSparkContext.<init>(JavaSparkContext.scala:61)
    at test.MLlib.BinarryClassification.main(BinaryClassification.java:105)

The numbers 41 and 105 are the lines, where both objects are defined in Java code. My question is, is it possible to execute multiple SparkContext on the same JVM and how to do it, if I already use the set -method ?

Answer 1

Are you sure you need the JavaSparkContext as a separate context? The previous question that you refer to doesn't say so. If you already have a Spark Context you can create a new JavaSparkContext from it, rather than create a separate context:

SparkConf conf = new SparkConf();
conf.setAppName("Spark MultipleContest Test");
conf.set("spark.driver.allowMultipleContexts", "true");
conf.setMaster("local");

SparkContext sc = new SparkContext(conf);
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

//Create a Java Context which is the same as the scala one under the hood
JavaSparkContext.fromSparkContext(sc)

Answer 2

默认情况下SparkContext正在运行，所以你必须停止这个上下文：sc.stop然后你可以继续没有任何pb

Answer 3

Rather than using SparkContext you should use builder method on SparkSession which more roubustly instantiates the spark and SQL context and ensures that there is not context conflict.

import org.apache.spark.sql.SparkSession
val spark = SparkSession.builder().appName("demo").getOrCreate()