大数据(十)---------Spark Streaming整合Kafka

最新推荐文章于 2024-09-11 14:10:29 发布

lpqupup

最新推荐文章于 2024-09-11 14:10:29 发布

阅读量138

点赞数 1

分类专栏： Spark 文章标签： Spark Spark Streaming 大数据 kafka

本文链接：https://blog.csdn.net/qq_40510501/article/details/96452510

版权

Spark 专栏收录该内容

1 篇文章 0 订阅

订阅专栏

kafka作为一个实时的分布式消息队列，实时的生产和消费消息，这里我们可以利用Spark Streaming实时计算框架实时地读取kafka中的数据然后进行计算

本次整合所需依赖：

<dependencies>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-streaming_2.11</artifactId>
        <version>2.0.2</version>
    </dependency>
    <dependency>
        <groupId>org.apache.spark</groupId>
        <artifactId>spark-streaming-kafka-0-8_2.11</artifactId>
        <version>2.1.0</version>
    </dependency>
</dependencies>

集群操作：

1.启动zookeeper集群

zkServer.sh start

2.启动kafka集群

kafka-server-start.sh  /opt/software/kafka/config/server.properties

3.创建topic

kafka-topics.sh --create --zookeeper node01:2181 --replication-factor 1 --partitions 3 --topic kafka_spark

4.向topic中生产数据

kafka-console-producer.sh --broker-list node01:9092 --topic  kafka_spark

编写Spark Streaming程序：

package com.nb.lpq

import org.apache.spark.streaming.dstream.DStream
import org.apache.spark.streaming.kafka.KafkaUtils
import org.apache.spark.streaming.{Seconds, StreamingContext}
import org.apache.spark.{SparkConf, SparkContext}

import scala.collection.immutable

//todo:利用sparkStreaming接受kafka中的数据实现单词计数----采用receivers
object SparkStreamingKafka_Receiver_checkpoint {
  def updateFunc(a:Seq[Int], b:Option[Int]) :Option[Int] ={
    Some(a.sum+b.getOrElse(0))
  }
  def main(args: Array[String]): Unit = {
    val checkpointPath = "./kafka-receiver"

    val ssc = StreamingContext.getOrCreate(checkpointPath, () => {
      createFunc(checkpointPath)
    })
    ssc.start()
    ssc.awaitTermination()
  }
  def createFunc(checkpointPath:String): StreamingContext = {

    //todo:1、创建sparkConf
    val sparkConf: SparkConf = new SparkConf()
      .setAppName("SparkStreamingKafka_Receiver_checkpoint")
      .setMaster("local[4]")
      //todo:开启wal预写日志
      .set("spark.streaming.receiver.writeAheadLog.enable","true")
    //todo:2、创建sparkContext
    val sc = new SparkContext(sparkConf)

    sc.setLogLevel("WARN")

    //todo:3、创建StreamingContext
    val ssc = new StreamingContext(sc,Seconds(5))
    ssc.checkpoint(checkpointPath)
    //todo:4、指定zkServer
    val zkServer="node02:2181,node03:2181,node04:2181"

    //todo:5、指定groupId
    val groupId="spark-kafka-receiver01"

    //todo:6、指定topics 这个可以利用一个消费者组来消费多个topic,
    //(topic_name -> numPartitions)  指定topic消费的线程数
    val topics=Map("kafka_spark"->1)

    //todo:7、并行运行更多的接收器读取kafak topic中的数据，这里设置3个
    val resultDStream: immutable.IndexedSeq[DStream[String]] = (1 to 3).map(x => {
      //todo:8、通过使用KafkaUtils的createStream接受kafka topic中的数据，生成DStream
      val kafkaDataDStream: DStream[String] = KafkaUtils.createStream(ssc, zkServer, groupId, topics).map(x => x._2)
      kafkaDataDStream
    }
    )
    //todo:利用StreamContext将所有的DStream组合在一起
    val kafkaDStream: DStream[String] = ssc.union(resultDStream)

    //todo:8、获取kafka中topic的内容

    //todo:9、切分每一行。每个单词记为1
    val wordAndOne: DStream[(String, Int)] = kafkaDStream.flatMap(_.split(" ")).map((_,1))

    //todo:10、相同单词出现的次数累加
    val result: DStream[(String, Int)] = wordAndOne.updateStateByKey(updateFunc)

    //todo:打印
    result.print()
    ssc

  }
}