Kafka Spark streaming: unable to read messages

6k views Asked by At

I am integrating Kafka and Spark, using spark-streaming. I have created a topic as a kafka producer:

bin/kafka-topics.sh --create --zookeeper localhost:2181 --replication-factor 1 --partitions 1 --topic test 

I am publishing messages in kafka and trying to read them using spark-streaming java code and displaying them on screen.
The daemons are all up: Spark-master,worker; zookeeper; kafka.
I am writing a java code for doing it, using KafkaUtils.createStream
code is below:

public class SparkStream {
    public static void main(String args[])
    {
        if(args.length != 3)
        {
            System.out.println("SparkStream <zookeeper_ip> <group_nm> <topic1,topic2,...>");
            System.exit(1);
        }


        Map<String,Integer> topicMap = new HashMap<String,Integer>();
        String[] topic = args[2].split(",");
        for(String t: topic)
        {
            topicMap.put(t, new Integer(1));
        }

        JavaStreamingContext jssc = new JavaStreamingContext("spark://192.168.88.130:7077", "SparkStream", new Duration(3000));
        JavaPairReceiverInputDStream<String, String> messages = KafkaUtils.createStream(jssc, args[0], args[1], topicMap );

        System.out.println("Connection done++++++++++++++");
        JavaDStream<String> data = messages.map(new Function<Tuple2<String, String>, String>() 
                                                {
                                                    public String call(Tuple2<String, String> message)
                                                    {
                                                        System.out.println("NewMessage: "+message._2()+"++++++++++++++++++");
                                                        return message._2();
                                                    }
                                                }
                                                );
        data.print();

        jssc.start();
        jssc.awaitTermination();

    }
}

I am running the job, and at other terminal I am running kafka-producer to publish messages:

Hi kafka
second message
another message

But the output logs at the spark-streaming console doesn't show the messages, but shows zero blocks received:

-------------------------------------------
Time: 1417438988000 ms
-------------------------------------------

2014-12-01 08:03:08,008 INFO  [sparkDriver-akka.actor.default-dispatcher-4] scheduler.JobScheduler (Logging.scala:logInfo(59)) - Starting job streaming job 1417438988000 ms.0 from job set of time 1417438988000 ms
2014-12-01 08:03:08,008 INFO  [sparkDriver-akka.actor.default-dispatcher-4] scheduler.JobScheduler (Logging.scala:logInfo(59)) - Finished job streaming job 1417438988000 ms.0 from job set of time 1417438988000 ms
2014-12-01 08:03:08,009 INFO  [sparkDriver-akka.actor.default-dispatcher-4] scheduler.JobScheduler (Logging.scala:logInfo(59)) - Total delay: 0.008 s for time 1417438988000 ms (execution: 0.000 s)
2014-12-01 08:03:08,010 INFO  [sparkDriver-akka.actor.default-dispatcher-15] scheduler.JobScheduler (Logging.scala:logInfo(59)) - Added jobs for time 1417438988000 ms
2014-12-01 08:03:08,015 INFO  [sparkDriver-akka.actor.default-dispatcher-15] rdd.MappedRDD (Logging.scala:logInfo(59)) - Removing RDD 39 from persistence list
2014-12-01 08:03:08,024 INFO  [sparkDriver-akka.actor.default-dispatcher-4] storage.BlockManager (Logging.scala:logInfo(59)) - Removing RDD 39
2014-12-01 08:03:08,027 INFO  [sparkDriver-akka.actor.default-dispatcher-15] rdd.BlockRDD (Logging.scala:logInfo(59)) - Removing RDD 38 from persistence list
2014-12-01 08:03:08,031 INFO  [sparkDriver-akka.actor.default-dispatcher-2] storage.BlockManager (Logging.scala:logInfo(59)) - Removing RDD 38
2014-12-01 08:03:08,033 INFO  [sparkDriver-akka.actor.default-dispatcher-15] kafka.KafkaInputDStream (Logging.scala:logInfo(59)) - Removing blocks of RDD BlockRDD[38] at BlockRDD at ReceiverInputDStream.scala:69 of time 1417438988000 ms
2014-12-01 08:03:09,002 INFO  [sparkDriver-akka.actor.default-dispatcher-2] scheduler.ReceiverTracker (Logging.scala:logInfo(59)) - Stream 0 received 0 blocks

Why isn't the data block getting received? i have tried using kafka producer-consumer on console bin/kafka-console-producer.... and bin/kafka-console-consumer... its working perfect, but why not my code... any idea ?

2

There are 2 answers

2
aiman On BEST ANSWER

Issue solved.

the code above is correct. We will just add two more lines to supress the [INFO] and [WARN] generated. So the final code is:

package com.spark;

import scala.Tuple2;
import org.apache.log4j.Logger;
import org.apache.log4j.Level;
import kafka.serializer.Decoder;
import kafka.serializer.Encoder;
import org.apache.spark.streaming.Duration;
import org.apache.spark.*;
import org.apache.spark.api.java.function.*;
import org.apache.spark.api.java.*;
import org.apache.spark.streaming.kafka.KafkaUtils;
import org.apache.spark.streaming.kafka.*;
import org.apache.spark.streaming.api.java.JavaStreamingContext;
import org.apache.spark.streaming.api.java.JavaPairDStream;
import org.apache.spark.streaming.api.java.JavaDStream;
import org.apache.spark.streaming.api.java.JavaPairReceiverInputDStream;
import java.util.Map;
import java.util.HashMap;

public class SparkStream {
    public static void main(String args[])
    {
        if(args.length != 3)
        {
            System.out.println("SparkStream <zookeeper_ip> <group_nm> <topic1,topic2,...>");
            System.exit(1);
        }

        Logger.getLogger("org").setLevel(Level.OFF);
        Logger.getLogger("akka").setLevel(Level.OFF);
        Map<String,Integer> topicMap = new HashMap<String,Integer>();
        String[] topic = args[2].split(",");
        for(String t: topic)
        {
            topicMap.put(t, new Integer(3));
        }

        JavaStreamingContext jssc = new JavaStreamingContext("local[4]", "SparkStream", new Duration(1000));
        JavaPairReceiverInputDStream<String, String> messages = KafkaUtils.createStream(jssc, args[0], args[1], topicMap );

        System.out.println("Connection done++++++++++++++");
        JavaDStream<String> data = messages.map(new Function<Tuple2<String, String>, String>() 
                                                {
                                                    public String call(Tuple2<String, String> message)
                                                    {
                                                        return message._2();
                                                    }
                                                }
                                                );
        data.print();

        jssc.start();
        jssc.awaitTermination();

    }
}

Also we need to add dependency in the POM.xml:

<dependency>
<groupId>com.msiops.footing</groupId>
<artifactId>footing-tuple</artifactId>
<version>0.2</version>
</dependency>  

This dependency is used for making use of the scala.Tuple2
The error of Stream 0 received 0 block was due to the spark-worker not available and and the spark-worker-core was set to 1. For spark-streaming we need the core to be >=2. So we need to make changes in the spark-config file. Refer the installation manual. to add the line export SPARK_WORKER_CORE=5 Also change the SPARK_MASTER='hostname' to SPARK_MASTER=<your local IP>. This local ip is what you see in BOLD when you go to your Spark UI web console...something like: spark://192.168..:<port>. We dont need the port here. only the IP is required.
Now restart your spark-master and spark-worker and start streaming :)

output:

-------------------------------------------
Time: 1417443060000 ms
-------------------------------------------
message 1

-------------------------------------------
Time: 1417443061000 ms
-------------------------------------------
message 2

-------------------------------------------
Time: 1417443063000 ms
-------------------------------------------
message 3
message 4

-------------------------------------------
Time: 1417443064000 ms
-------------------------------------------
message 5
message 6
messag 7

-------------------------------------------
Time: 1417443065000 ms
-------------------------------------------
message 8
4
Vijay Innamuri On

Yes, you need to access the content from DStream.

messages.foreachRDD(<<processing for the input received in the interval>>);