第17课：Scala并发编程实战及Spark源码阅读

本节课主题

1、Spark源码中并发编程源码阅读

2、Scala并发编程实战

早期Spark中是通过Akka进行并发通信，Spark 1.6.x以后是通过Netty，在Scala中是通过Actor来实现并发编程，就相当于收邮件和发邮件机制。

我们通过代码实战Actor

在进行Scala的Actor编程前（Scala 2.11.8版本），需要先引进2个jar包：

scala-actors.jar

scala-actors-migration.jar

参考示例代码：

package com.dt.scala
import scala.actors.Actor

case class Hello(name: String, content: String, sender: Actor)
case class HelloBack(name: String, content: String, sender: Actor)

class HelloActor extends Actor {
  var counter = 0
  override def act() {
    while (true) {
      receive {
        case content: String => println("Message:" + content)
        case Hello(name, content, sender) => {
          println("HelloActor:" + name + " content:" + content + counter)
          counter += 1
          Thread.sleep(3000) //睡眠3s
          sender ! HelloBack(name, content, this)  // Actor发送消息语法！
        }
      }
    }
  }
}

class HelloBackActor(val helloActor: Actor) extends Actor {
  var counter = 0
  override def act() {
    //启动给HelloActor发送信息
    helloActor ! Hello("Spark", "Love is here!", this) // this = HelloBackActor
    while (true) {
      receive {
        case content: String => println("Message:" + content)
        case HelloBack(name, content, sender) => {
          println("HelloBackActor:" + name + " content:" + content + counter)
          counter += 1
          Thread.sleep(3000) //睡眠3s
          sender ! Hello(name, content, this)
        }
      }
    }
  }
}

object HelloScalaActor {
  def main(args: Array[String]): Unit = {
    val helloActor = new HelloActor
    helloActor.start() //启动Actor来接收消息
    
    val hellBackActor = new HelloBackActor(helloActor)
    hellBackActor.start()

    //    var counter = 0
    //    while (true) {
    //      helloActor ! "Life is short,you need spark!!!" + counter //Actor发消息语法：!
    //      counter += 1
    //
    //      //发完消息睡眠3s
    //      Thread.sleep(3000)
    //    }

    //发送消息
    //helloActor ! Hello("Spark", "you need spark!!!")
  }
}

备注：

资料来源于：DT_大数据梦工厂（Spark"蘑菇云"行动）

更多私密内容，请关注微信公众号：DT_Spark

如果您对大数据Spark感兴趣，可以免费听由王家林老师每天晚上20：00开设的Spark永久免费公开课，

地址YY房间号：68917580