基于akka与scala实现一个简单rpc框架

本文主要是介绍基于akka与scala实现一个简单rpc框架，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

一、RPC简介

二、代码框架

三、代码实现

该小项目包含四个文件：

1.WorkerInfo，用于保存Worker的信息，此次保存Worker的上一次心跳时间

package com.zxl.rpcclass WorkerInfo(val id: String, val memory: Int, val cores: Int) {// TODO 上一次心跳var lastHeartbeatTime: Long = _
}

2.RemoteMessage，实现序列化并定义Master与Worker之间传送信息的类型

package com.zxl.rpc/*** 用于实现序列化 网络传输*/
trait RemoteMessage extends Serializable// Worker -> Master
case class RegisterWorker(id: String, memory: Int, cores: Int) extends RemoteMessage
case class Heartbeat(id: String) extends RemoteMessage// Master -> Worker
case class RegisteredWorker(masterUrl: String) extends RemoteMessage// Worker -> self
case object SendHeartbeat// Master -> self
case object CheckTimeOutWorker

3.Worker，与Master进行消息交互

package com.zxl.rpcimport java.util.UUID
import akka.actor.{Props, ActorSystem, Actor, ActorSelection}
import com.typesafe.config.ConfigFactory
import scala.concurrent.duration._class Worker(val masterHost: String, val masterPort: Int, val memory: Int, val cores: Int) extends Actor {// 与master连接的对象var master : ActorSelection = _// 每个worker的idval workerId = UUID.randomUUID().toString// 发送心跳的时间间隔val HEART_INTERVAL = 10000override def preStart(): Unit = {// 建立连接// 在Master启动时会打印下面的那个协议, 可以先用这个做一个标志, 连接哪个master// 继承actor后会有一个context, 可以通过它来连接// 需要有/user, Master要和master那边创建的名字保持一致master = context.actorSelection(s"akka.tcp://MasterSystem@$masterHost:$masterPort/user/Master")// 向Master发送注册消息master ! RegisterWorker(workerId, memory, cores)}override def receive: Receive = {case RegisteredWorker(masterUrl) => {println(masterUrl)// 启动定时器发送心跳import context.dispatcher// 多长时间后执行 单位,多长时间执行一次 单位, 消息的接受者(直接给master发不好, 先给自己发送消息, 以后可以做下判断, 什么情况下再发送消息), 信息context.system.scheduler.schedule(0 millis, HEART_INTERVAL millis, self, SendHeartbeat)}case SendHeartbeat => {println("send heartbeat to master")// 对master发送心跳信息，发送当前worker的idmaster ! Heartbeat(workerId)}}
}object Worker {def main(args: Array[String]) {val host = args(0)val port = args(1).toIntval masterHost = args(2)val masterPort = args(3).toInt// 分配的内存大小val memory = args(4).toInt// 分配的处理器核数val cores = args(5).toInt// 准备配置val configStr =s"""|akka.actor.provider = "akka.remote.RemoteActorRefProvider"|akka.remote.netty.tcp.hostname = "$host"|akka.remote.netty.tcp.port = "$port"""".stripMarginval config = ConfigFactory.parseString(configStr)// ActorSystem老大，辅助创建和监控下面的Actor，它是单例的val actorSystem = ActorSystem("WorkerSystem", config)actorSystem.actorOf(Props(new Worker(masterHost, masterPort, memory, cores)), "Worker")actorSystem.awaitTermination()}
}

4.Master，接收Worker的消息并做回应

package com.zxl.rpcimport akka.actor.{Props, ActorSystem, Actor}
import com.typesafe.config.ConfigFactory
import scala.collection.mutable
import scala.concurrent.duration._class Master(val host: String, val port: Int) extends Actor {// 保存(workerId,WorkerInfo)val idToWorker = new mutable.HashMap[String, WorkerInfo]()// 保存wokerInfo// 使用set删除快, 也可用linkListval workers = new mutable.HashSet[WorkerInfo]()// 超时检查的间隔val CHECK_INTERVAL = 15000override def preStart(): Unit = {println("preStart invoked")// 导入隐式转换import context.dispatcher// 使用timer太low了, 可以使用akka的, 使用定时器, 要导入这个包// 定时检查worker的心跳时间是否超时context.system.scheduler.schedule(0 millis, CHECK_INTERVAL millis, self, CheckTimeOutWorker)}// 用于接收消息override def receive: Receive = {case RegisterWorker(id, memory, cores) => {// 判断一下，是不是已经注册过if(!idToWorker.contains(id)) {// 把Worker的信息封装起来保存到内存当中val workerInfo = new WorkerInfo(id, memory, cores)idToWorker(id) = workerInfoworkers += workerInfo// 通知worker注册 并将master的地址返回给workersender ! RegisteredWorker(s"akka.tcp://MasterSystem@$host:$port/user/Master")}}case Heartbeat(id) => {if(idToWorker.contains(id)) {val workerInfo = idToWorker(id)// 报活val currentTime = System.currentTimeMillis()// 更新已收到id的worker的上一次心跳时间workerInfo.lastHeartbeatTime = currentTime}}case CheckTimeOutWorker => {val currentTime = System.currentTimeMillis()// 获取过时的workerval toRemove = workers.filter(x => currentTime - x.lastHeartbeatTime > CHECK_INTERVAL)// 将超时的worker从两个集合中去掉for(w <- toRemove) {workers -= widToWorker -= w.id}println(workers.size)}}
}object Master {def main(args: Array[String]) {val host = args(0)val port = args(1).toInt// 准备配置val configStr =s"""|akka.actor.provider = "akka.remote.RemoteActorRefProvider"|akka.remote.netty.tcp.hostname = "$host"|akka.remote.netty.tcp.port = "$port"""".stripMarginval config = ConfigFactory.parseString(configStr)// ActorSystem老大，辅助创建和监控下面的Actor，它是单例的val actorSystem = ActorSystem("MasterSystem", config)// 创建Actorval master = actorSystem.actorOf(Props(new Master(host, port)), "Master")master ! "hello"actorSystem.awaitTermination()}
}

这篇关于基于akka与scala实现一个简单rpc框架的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！

---