Note: this post was revised in January 20th, 2017 to reflect changes in Kafka
Background:
When using Apache Kafka, one concern raised is how to run unit tests for the consumers without the need to start a whole Kafka cluster and Zookeeper.
In Kafka 0.9 two Mock classes was added: MockProducer and MockConsumer.
The problem with those mock classes is that in many cases they are just unusable. The reason is that we use frameworks for Kafka consumers that do not allow to implement a mock class instead of the real KafkaConsumer that is internally used.
But still, we want to be able to test our code somehow without the need to start Zookeeper and Kafka server always.
In this article I suggest a new approach that can be helpful in case you want to write a unit-test for your code that implements Kafka Consumer.
Test case
I take for example the case of Spark Streaming using Kafka Receiver.
The whole example can be found in this GitHub repository.
In this project, I set a code example for Spark Streaming using Kafka receiver to perform word count application. There is code in both Java and Scala.
This is the Word Count Java code to be tested (complete code can be found here):
package com.myspark;
import com.google.common.collect.Lists;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.*;
import org.apache.spark.streaming.Duration;
import org.apache.spark.streaming.api.java.*;
import org.apache.spark.streaming.kafka.KafkaUtils;
import scala.Tuple2;
import java.util.HashMap;
import java.util.Map;
import java.util.regex.Pattern;
public final class JavaKafkaWordCount {
private static final Pattern SPACE = Pattern.compile(" ");
private JavaKafkaWordCount() {
}
public static void main(String[] args) {
if (args.length < 4) {
System.err.println("Usage: JavaKafkaWordCount <zkQuorum> <group> <topics> <numThreads>");
System.exit(1);
}
LoggerTools.setStreamingLogLevels();
SparkConf sparkConf = new SparkConf().setAppName("JavaKafkaWordCount");
sparkConf.setMaster("local[2]"); //set master server
sparkConf.set("com.couchbase.bucket.travel-sample", "");
// Create the context with 2 seconds batch size
JavaStreamingContext jssc = new JavaStreamingContext(sparkConf, new Duration(2000));
int numThreads = Integer.parseInt(args[3]);
Map<String, Integer> topicMap = new HashMap<String, Integer>();
String[] topics = args[2].split(",");
for (String topic: topics) {
topicMap.put(topic, numThreads);
}
JavaPairReceiverInputDStream<String, String> messages =
KafkaUtils.createStream(jssc, args[0], args[1], topicMap);
JavaDStream<String> lines = messages.map(new Function<Tuple2<String, String>, String>() {
@Override
public String call(Tuple2<String, String> tuple2) {
return tuple2._2();
}
});
JavaDStream<String> words = lines.flatMap(new FlatMapFunction<String, String>() {
@Override
public Iterable<String> call(String x) {
return Lists.newArrayList(SPACE.split(x));
}
});
JavaPairDStream<String, Integer> wordCounts = words.mapToPair(
new PairFunction<String, String, Integer>() {
@Override
public Tuple2<String, Integer> call(String s) {
return new Tuple2<String, Integer>(s, 1);
}
}).reduceByKey(new Function2<Integer, Integer, Integer>() {
@Override
public Integer call(Integer i1, Integer i2) {
return i1 + i2;
}
});
wordCounts.print();
jssc.start();
jssc.awaitTermination();
}
}
And now – the test
In order to test the class above, we perform the following steps:
- Start a local Zookeeper server
- Start a local Kafka server
- Create Kafka Producer
- Run the Spark Streaming program
- Send some messages through the Kafka Producer
The code below does this (complete code can be found here):
package com.myspark;
import kafka.server.KafkaConfig;
import kafka.server.KafkaServerStartable;
import org.apache.kafka.clients.producer.*;
import org.apache.zookeeper.server.ZooKeeperServerMain;
import org.junit.*;
import java.io.File;
import java.io.IOException;
import java.nio.charset.Charset;
import java.util.Properties;
public class JavaKafkaWordCountTest {
private static final String TOPIC = "topic-1";
private static final String BROKERHOST = "127.0.0.1";
private static final String BROKERPORT = "9092";
private static final String ZKPORT = "2181";
private String nodeId = "0";
private String zkConnect = "localhost:" + ZKPORT;
private KafkaServerStartable server;
KafkaProducer<Integer, byte[]> producer;
@Before
public void setup() throws IOException {
//zookeeper
startZK();
//start kafka
startKafka();
// setup producer
setupProducer();
}
@After
public void tearDown() throws Exception {
server.shutdown();
server.awaitShutdown();
}
private static void startZK() throws IOException {
final File zkTmpDir = File.createTempFile("zookeeper", "test");
zkTmpDir.delete();
zkTmpDir.mkdir();
new Thread() {
@Override
public void run() {
ZooKeeperServerMain.main(new String [] {ZKPORT, zkTmpDir.getAbsolutePath()});
}
}.start();
try {
Thread.sleep(1000);
} catch (InterruptedException e) {
}
}
private void startKafka() {
Properties props = new Properties();
props.put("broker.id", nodeId);
props.put("port", BROKERPORT);
props.put("zookeeper.connect", zkConnect);
props.put("host.name", "127.0.0.1");
KafkaConfig conf = new KafkaConfig(props);
server = new KafkaServerStartable(conf);
server.startup();
}
private void setupProducer() {
Properties producerProps = new Properties();
producerProps.setProperty("bootstrap.servers", BROKERHOST + ":" + BROKERPORT);
producerProps.setProperty("key.serializer","org.apache.kafka.common.serialization.IntegerSerializer");
producerProps.setProperty("value.serializer", "org.apache.kafka.common.serialization.ByteArraySerializer");
producer = new KafkaProducer<>(producerProps);
}
@Test
public void testSparkWordCount() throws Exception {
Thread t = new Thread(() -> {
String[] args = {"localhost", "grp-1", TOPIC, "2"};
JavaKafkaWordCount.main(args);
System.out.println("End Child Thread");
});
t.start();
for (int i=0; i<1000; i++){
producer.send(new ProducerRecord<>(TOPIC, 0, 1, ("There are some words here to count -" + Integer.toString(i)).getBytes(Charset.forName("UTF-8"))));
Thread.sleep(10);
}
System.out.println("End Test");
}
}
Ran Silberman 