spark 源码解析其一(资源的申请,任务调度)

一,资源的申请 以下来自于 Master
    1,首先receiveAndReply 方法接收EndpointRef ask及其衍生方法发送的信息(模式匹配RequestSubmitDriver)
      1),创建 driver 信息
      2),driver 持久化 , 存储 driver 运行记录, 以便于恢复信息
      3), 调用 schedule() 检查并配置 executor信息, 提交 job

  override def receiveAndReply(context: RpcCallContext): PartialFunction[Any, Unit] = {
    case RequestSubmitDriver(description) =>
      if (state != RecoveryState.ALIVE) {
        val msg = s"${Utils.BACKUP_STANDALONE_MASTER_PREFIX}: $state. " +
          "Can only accept driver submissions in ALIVE state."
        context.reply(SubmitDriverResponse(self, false, None, msg))
      } else {
        logInfo("Driver submitted " + description.command.mainClass)
        // 创建 driver 信息
        val driver = createDriver(description)
//        driver 持久化 , 存储 driver 运行记录, 以便于恢复信息
        persistenceEngine.addDriver(driver)

        waitingDrivers += driver
        drivers.add(driver)
//        调度 driver
        schedule()

        // TODO: It might be good to instead have the submission client poll the master to determine
        //       the current status of the driver. For now it's simply "fire and forget".
//        提交 Driver 资源
        context.reply(SubmitDriverResponse(self, true, Some(driver.id),
          s"Driver successfully submitted as ${driver.id}"))
      }

View Code

　　2,schedule 配置文件获取信息
  　　1),打散集群 所有的 worker
  　　2),配置文件读取 driver memory 与 core
  　　3),创建 executor

  private def schedule(): Unit = {
//    判断 driver 是不是  Alive
    if (state != RecoveryState.ALIVE) {
      return
    }
    // Drivers take strict precedence over executors
//    打散集群 所有的 worker
    val shuffledAliveWorkers = Random.shuffle(workers.toSeq.filter(_.state == WorkerState.ALIVE))
    val numWorkersAlive = shuffledAliveWorkers.size
    var curPos = 0
    for (driver <- waitingDrivers.toList) { // iterate over a copy of waitingDrivers
      // We assign workers to each waiting driver in a round-robin fashion. For each driver, we
      // start from the last worker that was assigned a driver, and continue onwards until we have
      // explored all alive workers.
      var launched = false
      var numWorkersVisited = 0
      while (numWorkersVisited < numWorkersAlive && !launched) {
        val worker = shuffledAliveWorkers(curPos)
        numWorkersVisited += 1
//         判读 driver memory 与 core
        if (worker.memoryFree >= driver.desc.mem && worker.coresFree >= driver.desc.cores) {
//           启动 driver work: driver 所要运行的 位置
//          driver : 当前 driver 的信息, SparkContext(各种任务调度器)
          launchDriver(worker, driver)
          waitingDrivers -= driver
          launched = true
        }
        curPos = (curPos + 1) % numWorkersAlive
      }
    }
//    给 job 分配资源, 对于 job 来说 ,资源指的是 Executor
    startExecutorsOnWorkers()
  }

View Code

　　3,startExecutorsOnWorkers
  　　1),遍历集群之中所有的满足 app 要求的 core 和 memory 的 work 信息,并返回给Schedule(),在此返回给 RequestSubmitDriver 提交job

  private def startExecutorsOnWorkers(): Unit = {
    // Right now this is a very simple FIFO scheduler. We keep trying to fit in the first app
    // in the queue, then the second app, etc.
//    waitingApps 代表的是 所有 master 注册的 作业(Executors)
    for (app <- waitingApps) {
      val coresPerExecutor = app.desc.coresPerExecutor.getOrElse(1)
      // If the cores left is less than the coresPerExecutor,the cores left will not be allocated
//      app.coresLeft = --toal-excutor-cores - 分配的 core
//      coresPerExecutor = --executor-cores(每个 excutor 使用的 core)
      if (app.coresLeft >= coresPerExecutor) {
        // Filter out workers that don't have enough resources to launch an executor
//       work 指的是 集群所有的 worker 信息
//        遍历集群之中所有的满足 app 要求的 core 和 memory 的 work 信息,剩余的 core 多到少之中排序
        val usableWorkers = workers.toArray.filter(_.state == WorkerState.ALIVE)
          .filter(worker => worker.memoryFree >= app.desc.memoryPerExecutorMB &&
            worker.coresFree >= coresPerExecutor)
          .sortBy(_.coresFree).reverse
//    每个 worker 上可以为 当前 application 提供的 core 数
        val assignedCores = scheduleExecutorsOnWorkers(app, usableWorkers, spreadOutApps)

        // Now that we've decided how many cores to allocate on each worker, let's allocate them
        for (pos <- 0 until usableWorkers.length if assignedCores(pos) > 0) {
          allocateWorkerResourceToExecutors(
            app, assignedCores(pos), app.desc.coresPerExecutor, usableWorkers(pos))
        }
      }
    }
  }

View Code

二,任务调度源码解析(DAGScheduler 类), 主要是(getMissingParentStages()与submitStage()两个方法):
    1,首先从 foreach() 开始, 根据最后一个 RDD 创建 finalStage
    2, 切割 RDD 划分 stage(getMissingParentStages())
    1),如果failedStages的父stage是宽依赖,就存到 missing:HashSet
    2),如果failedStages的父stage是窄依赖, 那么就会压栈,然后一直遍历栈,直至划分完最后的 RDD
    3),getMissingParentStages() 最终 返回划分好的 stage 也就是  missing.toList
    3,submitMissingTasks() 之中, 根据 missing（每stage划分好的job） 递归提交所有的job

  // 切割 RDD 划分 stage
  //  递归方法 递归提交父 stage,从 finalStage 开始
  // finalStage = createResultStage(finalRDD, func, partitions, jobId, callSite)
  private def getMissingParentStages(stage: Stage): List[Stage] = {
    val missing = new HashSet[Stage]
    val visited = new HashSet[RDD[_]]
    // We are manually maintaining a stack here to prevent StackOverflowError
    // caused by recursively visiting
    val waitingForVisit = new ArrayStack[RDD[_]]
    def visit(rdd: RDD[_]) {
      if (!visited(rdd)) {
        visited += rdd
        val rddHasUncachedPartitions = getCacheLocs(rdd).contains(Nil)
        if (rddHasUncachedPartitions) {
          for (dep <- rdd.dependencies) {
            dep match {
//                窄依赖
              case shufDep: ShuffleDependency[_, _, _] =>
//                创建 stage
                val mapStage = getOrCreateShuffleMapStage(shufDep, stage.firstJobId)
                if (!mapStage.isAvailable) {
//                  // 将 stage 放置  missing 之中
                  missing += mapStage
                }
//                宽依赖
              case narrowDep: NarrowDependency[_] =>
                waitingForVisit.push(narrowDep.rdd)
            }
          }
        }
      }
    }
    waitingForVisit.push(stage.rdd)
    while (waitingForVisit.nonEmpty) {
      visit(waitingForVisit.pop())
    }
//    // 最后返回划分好的 stage
    missing.toList
  }



  /** Submits stage, but first recursively submits any missing parents. */
//  递归方法 递归提交父 stage
  private def submitStage(stage: Stage) {
    val jobId = activeJobForStage(stage)
    if (jobId.isDefined) {
      logDebug("submitStage(" + stage + ")")
      if (!waitingStages(stage) && !runningStages(stage) && !failedStages(stage)) {
//        寻找 丢失的父 stage
        val missing = getMissingParentStages(stage).sortBy(_.id)
        logDebug("missing: " + missing)
        if (missing.isEmpty) {
          logInfo("Submitting " + stage + " (" + stage.rdd + "), which has no missing parents")
//          递归完成所有的
          submitMissingTasks(stage, jobId.get)
        } else {
          for (parent <- missing) {
//            递归
            submitStage(parent)
          }
          waitingStages += stage
        }
      }
    } else {
      abortStage(stage, "No active job for stage " + stage.id, None)
    }
  }

View Code

 

转载于:https://www.cnblogs.com/shi-qi/articles/11254559.html