Kafka Spark 流式传输：无法读取消息

Question

我正在使用 spark-streaming 集成 Kafka 和 Spark。我作为 kafka 制作人创建了一个主题：

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

我在 kafka 中发布消息并尝试使用 spark-streaming java 代码读取它们并将它们显示在屏幕上。
守护进程全部启动：Spark-master,worker；动物园管理员；卡夫卡。
我正在编写一个 java 代码，使用 KafkaUtils.createStream
代码如下：

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();

    }
}

我正在运行作业，而在其他终端我正在运行 kafka-producer 来发布消息：

Hi kafka
second message
another message

但是 spark-streaming 控制台的输出日志不显示消息，但显示接收到零块：

-------------------------------------------
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

为什么没有收到数据块？我曾尝试在控制台bin/kafka-console-producer.... 和bin/kafka-console-consumer... 上使用kafka producer-consumer，它的工作完美，但为什么不是我的代码......任何想法？

Answer 1

问题已解决。

上面的代码是正确的。 我们将只添加两行来抑制生成的 [INFO] 和 [WARN]。所以最终的代码是：

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();

    }
}

另外我们需要在 POM.xml 中添加依赖：

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

此依赖项用于使用scala.Tuple2
Stream 0 received 0 block 的错误是由于 spark-worker 不可用并且 spark-worker-core 设置为 1。对于 spark-streaming，我们需要核心 >=2。 所以我们需要在 spark-config 文件中进行更改。请参阅安装手册。添加行export SPARK_WORKER_CORE=5 还将SPARK_MASTER='hostname' 更改为SPARK_MASTER=<your local IP>。当您转到 Spark UI Web 控制台时，此本地 IP 以粗体显示...类似于：spark://192.168..:<port>。我们不需要这里的端口。只需要 IP。
现在重新启动您的 spark-master 和 spark-worker 并开始流式传输:)

输出：

-------------------------------------------
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

Answer 2

是的，您需要从 DStream 访问内容。

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