Spark 资源调度包 stage 类解析

Posted 2020-11-16 ronnieyuan

spark 资源调度包 Stage(阶段) 类解析

• 类注释:

  /**
   * A stage is a set of parallel tasks all computing the same function that need to run as part
   * of a Spark job, where all the tasks have the same shuffle dependencies. 
   * 一个阶段是所有计算相同功能的并行任务集合, 作为spark作业的一部分, 这些任务都有相同的 shuffle 依赖
   *
   * Each DAG of tasks run by the scheduler is split up into stages at the boundaries where  
   * shuffle occurs, and then the DAGScheduler runs these stages in topological order.
   * 每个由调度器运行的任务的有向无环图在shuffle发生的分界处分化成不同的阶段, 并且这些有向无环图的调度器 
   * 将以拓扑排序来运行这些 不同阶段的任务
   *
   * Each Stage can either be a shuffle map stage, in which case its tasks' results are input for
   * other stage(s), or a result stage, in which case its tasks directly compute a Spark action
   * (e.g. count(), save(), etc) by running a function on an RDD.
   * 每个阶段都可以是一个 shuffle 匹配阶段, 该任务的结果可以被其他阶段作为导入, 也可以是 一个结果阶段,  
   * 它的任务 通过 在RDD上运行 一个公式 来直接计算 一个 spark 行为(计数, 保存等...)
   *
   * For shuffle map stages, we also track the nodes that each output partition is on.
   * 对于 shuffle 匹配阶段, 我们会追踪 每个输出的分区的节点
   *
   * Each Stage also has a firstJobId, identifying the job that first submitted the stage.  When 
   * FIFO scheduling is used, this allows Stages from earlier jobs to be computed first or 
   * recovered faster on failure.
   * 每个阶段都有一个 首任务Id, 用于辨识第一个提交到该阶段的 任务。 当 先进先出 调度策略被使用时, 这会允
   * 许 更早 的 作业 的阶段先被计算, 或者 在失败之后更早的恢复。
   *
   * Finally, a single stage can be re-executed in multiple attempts due to fault recovery. In 
   * that case, the Stage object will track multiple StageInfo objects to pass to listeners or  * the web UI.
   * 最终, 由于 容错恢复机制, 单个阶段可以在多次尝试运行中 被重新执行。在这种情况下, 该阶段对象会追踪 许
   * 多 阶段信息对象 并 将信息传递给监听者 或者 web UI 界面
   *
   * The latest one will be accessible through latestInfo.
   * 最新的阶段可以通过 最新的信息获取
   *
   * @param id Unique stage ID    id: 唯一的阶段Id
   * @param rdd RDD that this stage runs on: for a shuffle map stage, it's the RDD we run map 
   * tasks on, while for a result stage, it's the target RDD that we ran an action on
   * rdd：该 阶段 所运行在的 RDD：对于shuffle匹配阶段, 该RDD是我们 运行匹配任务 所在的RDD, 但对于结果阶
   * 段，该RDD是我们运行 行为所在的RDD
   * @param numTasks Total number of tasks in stage; 当前阶段的任务数量
   * result stages in particular may not need to compute all partitions, e.g. for first(), 
   * lookup(), and take().
   * 结果阶段 实际上不一定需要计算所有分区, 比如使用 first(), lookup() 和 take() 等算子
   * @param parents List of stages that this stage depends on (through shuffle dependencies).
   * parent: 该阶段 所依赖(shuffle依赖) 的 阶段列表
   * @param firstJobId ID of the first job this stage was part of, for FIFO scheduling.
   * firstJobId: 该阶段的第一个任务Id, 用于先进先出调度
   * @param callSite Location in the user program associated with this stage: either where the 
   * target RDD was created, for a shuffle map stage, or where the action for a result stage was 
   * called.
   * callSite: 与该阶段相关的用户程序的位置: 对于 shuffle 匹配阶段 是 目标 RDD创建的位置, 对于 结果
   * 阶段, 是 行为被调用的位置 （调用位点)
   */

• 代码:

  private[scheduler] abstract class Stage(
      val id: Int,
      val rdd: RDD[_],
      val numTasks: Int,
      val parents: List[Stage],
      val firstJobId: Int,
      val callSite: CallSite)
    extends Logging {
    // 根据数组长度确定分区数
    val numPartitions = rdd.partitions.length
  
    /** Set of jobs that this stage belongs to. 该阶段 的任务Id集合*/
    val jobIds = new HashSet[Int]
  
    /** The ID to use for the next new attempt for this stage. 
     * 该阶段的下一次新尝试的 Id
     */
    private var nextAttemptId: Int = 0
    // 短型的调用位点
    val name: String = callSite.shortForm
    // 长型的调用位点    
    val details: String = callSite.longForm
  
    /**
     * Pointer to the [[StageInfo]] object for the most recent attempt. 
     * 指向阶段信息对象的指针, 用于获取最近最多的尝试
     * This needs to be initialized here, before any attempts have actually been created, because 
     * the DAGScheduler uses this StageInfo to tell SparkListeners when a job starts (which 
     * happens before any stage attempts have been created).
     * 在任何尝试被实际创建之前, 它需要被初始化, 因为 有向无环图的调度器 使用了该阶段信息来告诉spark的监
     * 听者有任务启动了(这发生在任何阶段的尝试被发生之前)
     */
    private var _latestInfo: StageInfo = StageInfo.fromStage(this, nextAttemptId)
  
    /**
     * Set of stage attempt IDs that have failed with a FetchFailure. 
     * 拉取失败的的阶段尝试Id集
     * We keep track of these failures in order to avoid endless retries if a stage keeps failing 
     * with a FetchFailure.
     * 我们追踪这些失败时为了防止 一个阶段一直发生拉取失败后 无尽的重试
     * We keep track of each attempt ID that has failed to avoid recording duplicate failures if
     * multiple tasks from the same stage attempt fail (SPARK-5945).
     * 我们追踪失败的每次尝试的Id以防止 如果许多任务在同一个阶段的尝试中失败 会导致 记录重复的失败
     */
    val fetchFailedAttemptIds = new HashSet[Int]
    // 清除失败Id
    private[scheduler] def clearFailures() : Unit = {
      fetchFailedAttemptIds.clear()
    }
  
    /** Creates a new attempt for this stage by creating a new StageInfo with a new attempt ID.   
     * 通过创建一个新的附带新的尝试Id的状态信息 来为当前阶段创建一个新的尝试
     */
    def makeNewStageAttempt(
        // 需要计算的分区
        numPartitionsToCompute: Int,
        // 本地优先的任务
        taskLocalityPreferences: Seq[Seq[TaskLocation]] = Seq.empty): Unit = {
        // TaskMetrics是内部累加器的一个包装类
      val metrics = new TaskMetrics
      metrics.register(rdd.sparkContext)
       // 获取最近的阶段信息 
      _latestInfo = StageInfo.fromStage(
        this, nextAttemptId, Some(numPartitionsToCompute), metrics, taskLocalityPreferences)
      nextAttemptId += 1
    }
  
    /** Returns the StageInfo for the most recent attempt for this stage. 
     * 返回最近最多尝试的 当前阶段的 阶段信息
     */
    def latestInfo: StageInfo = _latestInfo
  
    override final def hashCode(): Int = id
  
    override final def equals(other: Any): Boolean = other match {
      case stage: Stage => stage != null && stage.id == id
      case _ => false
    }
  
    /** Returns the sequence of partition ids that are missing (i.e. needs to be computed). 
     * 返回缺失的分区 id 的序列
     */
    def findMissingPartitions(): Seq[Int]
  }