CC00088.spark——|Hadoop&Spark.V04|——|Spark.v04|Spark 原理 源码|作业执行原理&Stage划分|

一、stage划分

### --- dagScheduler.submit 发送消息

~~~     # 源码提取说明：DAGScheduler.scala
~~~     # 676行~703行
  def submitJob[T, U](
    rdd: RDD[T],
    func: (TaskContext, Iterator[T]) => U,
    partitions: Seq[Int],
    callSite: CallSite,
    resultHandler: (Int, U) => Unit,
    properties: Properties): JobWaiter[U] = {
    // Check to make sure we are not launching a task on a partition that does not exist.
    // 获取当前Job的最大分区数
    val maxPartitions = rdd.partitions.length
    // 检查不存在的分区，如果有就抛出异常
    partitions.find(p => p >= maxPartitions || p < 0).foreach { p =>
      throw new IllegalArgumentException(
        "Attempting to access a non-existent partition: " + p + ". " +
          "Total number of partitions: " + maxPartitions)
    }
    // 生成下一个Job的jobId
    val jobId = nextJobId.getAndIncrement()
    /**
     * 如果Job的分区数量等于0，则创建一个totalTasks属性为0的JobWaiter并返回。
     * 根据JobWaiter的实现，totalTasks属性为0的JobWaiter的jobPromise将被设置为Success。
     */
    if (partitions.size == 0) {
      // Return immediately if the job is running 0 tasks
      return new JobWaiter[U](this, jobId, 0, resultHandler)
    }
    assert(partitions.size > 0)
    // 分区数量大于0
    val func2 = func.asInstanceOf[(TaskContext, Iterator[_]) => _]
    // 创建JobWaiter
    val waiter = new JobWaiter(this, jobId, partitions.size, resultHandler)
    /**
     * 将JobWaiter包装到JobSubmitted消息中，投递给DAGSchedulerEventProcessLoop，
     * 这个消息最终会被DAGScheduler的handleJobSubmitted()方法处理。
     */
    eventProcessLoop.post(JobSubmitted(
      jobId, rdd, func2, partitions.toArray, callSite, waiter,
      SerializationUtils.clone(properties)))
    // 返回JobWaiter
    waiter
  }