DStream与RDD关系是什么

* DStreams internally is characterized by a few basic properties:
*  - A list of other DStreams that the DStream depends on
*  - A time interval at which the DStream generates an RDD
*  - A function that is used to generate an RDD after each time interval

// RDDs generated, marked as private[streaming] so that testsuites can access it
@transient
private[streaming] var generatedRDDs = new HashMap[Time, RDD[T]] ()

/**
 * Get the RDD corresponding to the given time; either retrieve it from cache
 * or compute-and-cache it.
 */
private[streaming] final def getOrCompute(time: Time): Option[RDD[T]] = {
  // If RDD was already generated, then retrieve it from HashMap,
  // or else compute the RDD
  generatedRDDs.get(time).orElse {
    // Compute the RDD if time is valid (e.g. correct time in a sliding window)
    // of RDD generation, else generate nothing.
    if (isTimeValid(time)) {

      val rddOption = createRDDWithLocalProperties(time, displayInnerRDDOps = false) {
        // Disable checks for existing output directories in jobs launched by the streaming
        // scheduler, since we may need to write output to an existing directory during checkpoint
        // recovery; see SPARK-4835 for more details. We need to have this call here because
        // compute() might cause Spark jobs to be launched.
        PairRDDFunctions.disableOutputSpecValidation.withValue(true) {
          compute(time)
        }
      }

      rddOption.foreach { case newRDD =>
        // Register the generated RDD for caching and checkpointing
        if (storageLevel != StorageLevel.NONE) {
          newRDD.persist(storageLevel)
          logDebug(s"Persisting RDD ${newRDD.id} for time $time to $storageLevel")
        }
        if (checkpointDuration != null && (time - zeroTime).isMultipleOf(checkpointDuration)) {
          newRDD.checkpoint()
          logInfo(s"Marking RDD ${newRDD.id} for time $time for checkpointing")
        }
        generatedRDDs.put(time, newRDD)
      }
      rddOption
    } else {
      None
    }
  }
}

/**
 * Generates RDDs with blocks received by the receiver of this stream. */
override def compute(validTime: Time): Option[RDD[T]] = {
  val blockRDD = {
    if (validTime < graph.startTime) {
      // If this is called for any time before the start time of the context,
      // then this returns an empty RDD. This may happen when recovering from a
      // driver failure without any write ahead log to recover pre-failure data.
      new BlockRDD[T](ssc.sc, Array.empty)
    } else {
      // Otherwise, ask the tracker for all the blocks that have been allocated to this stream
      // for this batch
      val receiverTracker = ssc.scheduler.receiverTracker
      val blockInfos = receiverTracker.getBlocksOfBatch(validTime).getOrElse(id, Seq.empty)

      // Register the input blocks information into InputInfoTracker
      val inputInfo = StreamInputInfo(id, blockInfos.flatMap(_.numRecords).sum)
      ssc.scheduler.inputInfoTracker.reportInfo(validTime, inputInfo)

      // Create the BlockRDD
      createBlockRDD(validTime, blockInfos)
    }
  }
  Some(blockRDD)
}

private[streaming] def createBlockRDD(time: Time, blockInfos: Seq[ReceivedBlockInfo]): RDD[T] = {
  if (blockInfos.nonEmpty) {
    val blockIds = blockInfos.map { _.blockId.asInstanceOf[BlockId] }.toArray

    // Are WAL record handles present with all the blocks
    val areWALRecordHandlesPresent = blockInfos.forall { _.walRecordHandleOption.nonEmpty }

    if (areWALRecordHandlesPresent) {
      // If all the blocks have WAL record handle, then create a WALBackedBlockRDD
      val isBlockIdValid = blockInfos.map { _.isBlockIdValid() }.toArray
      val walRecordHandles = blockInfos.map { _.walRecordHandleOption.get }.toArray
      new WriteAheadLogBackedBlockRDD[T](
        ssc.sparkContext, blockIds, walRecordHandles, isBlockIdValid)
    } else {
      // Else, create a BlockRDD. However, if there are some blocks with WAL info but not
      // others then that is unexpected and log a warning accordingly.
      if (blockInfos.find(_.walRecordHandleOption.nonEmpty).nonEmpty) {
        if (WriteAheadLogUtils.enableReceiverLog(ssc.conf)) {
          logError("Some blocks do not have Write Ahead Log information; " +
            "this is unexpected and data may not be recoverable after driver failures")
        } else {
          logWarning("Some blocks have Write Ahead Log information; this is unexpected")
        }
      }
      val validBlockIds = blockIds.filter { id =>
        ssc.sparkContext.env.blockManager.master.contains(id)
      }
      if (validBlockIds.size != blockIds.size) {
        logWarning("Some blocks could not be recovered as they were not found in memory. " +
          "To prevent such data loss, enabled Write Ahead Log (see programming guide " +
          "for more details.")
      }
      new BlockRDD[T](ssc.sc, validBlockIds)
    }
  } else {
    // If no block is ready now, creating WriteAheadLogBackedBlockRDD or BlockRDD
    // according to the configuration
    if (WriteAheadLogUtils.enableReceiverLog(ssc.conf)) {
      new WriteAheadLogBackedBlockRDD[T](
        ssc.sparkContext, Array.empty, Array.empty, Array.empty)
    } else {
      new BlockRDD[T](ssc.sc, Array.empty)
    }
  }
}

private[streaming]
class MappedDStream[T: ClassTag, U: ClassTag] (
    parent: DStream[T],
    mapFunc: T => U
  ) extends DStream[U](parent.ssc) {

  override def dependencies: List[DStream[_]] = List(parent)

  override def slideDuration: Duration = parent.slideDuration

  override def compute(validTime: Time): Option[RDD[U]] = {
    parent.getOrCompute(validTime).map(_.map[U](mapFunc))
  }
}

private[streaming]
class ForEachDStream[T: ClassTag] (
    parent: DStream[T],
    foreachFunc: (RDD[T], Time) => Unit,
    displayInnerRDDOps: Boolean
  ) extends DStream[Unit](parent.ssc) {

  override def dependencies: List[DStream[_]] = List(parent)

  override def slideDuration: Duration = parent.slideDuration

  override def compute(validTime: Time): Option[RDD[Unit]] = None

  override def generateJob(time: Time): Option[Job] = {
    parent.getOrCompute(time) match {
      case Some(rdd) =>
        val jobFunc = () => createRDDWithLocalProperties(time, displayInnerRDDOps) {
          foreachFunc(rdd, time)
        }
        Some(new Job(time, jobFunc))
      case None => None
    }
  }
}

/** Generate jobs and perform checkpoint for the given `time`.  */
private def generateJobs(time: Time) {
  // Set the SparkEnv in this thread, so that job generation code can access the environment
  // Example: BlockRDDs are created in this thread, and it needs to access BlockManager
  // Update: This is probably redundant after threadlocal stuff in SparkEnv has been removed.
  SparkEnv.set(ssc.env)
  Try {
    //根据特定的时间获取具体的数据
    jobScheduler.receiverTracker.allocateBlocksToBatch(time) // allocate received blocks to batch
    //调用DStreamGraph的generateJobs生成Job
    graph.generateJobs(time) // generate jobs using allocated block
  } match {
    case Success(jobs) =>
      val streamIdToInputInfos = jobScheduler.inputInfoTracker.getInfo(time)
      jobScheduler.submitJobSet(JobSet(time, jobs, streamIdToInputInfos))
    case Failure(e) =>
      jobScheduler.reportError("Error generating jobs for time " + time, e)
  }
  eventLoop.post(DoCheckpoint(time, clearCheckpointDataLater = false))
}

def generateJobs(time: Time): Seq[Job] = {
  logDebug("Generating jobs for time " + time)
  val jobs = this.synchronized {
    outputStreams.flatMap { outputStream =>
      val jobOption = outputStream.generateJob(time)
      jobOption.foreach(_.setCallSite(outputStream.creationSite))
      jobOption
    }
  }
  logDebug("Generated " + jobs.length + " jobs for time " + time)
  jobs
}