摘要
spark的调度一直是我想搞清楚的东西，以及有向无环图的生成过程、task的调度、rdd的延迟执行是怎么发生的和如何完成的，还要就是RDD的compute都是在executor的哪个阶段调用和执行我们定义的函数的。这些都非常的基础和困难。花一段时间终于弄白了其中的奥秘。总结起来，以便以后继续完善。spark的调度分为两级调度：DAGSchedule和TaskSchedule。DAGSchedule是根据job来生成相互依赖的stages，然后把stages以TaskSet形式传递给TaskSchedule来进行任务的分发过程，里面的细节会慢慢的讲解出来的，比较长。
本文目录
1、spark的RDD逻辑执行链
2、spark的job的划分、stage的划分
3、spark的DAGScheduler的调度
4、spark的TaskSchedule的调度
5、executor如何执行task以及我们定义的函数
spark的RDD的逻辑执行链
都说spark进行延迟执行，通过RDD的DAG来生成相应的Stage等，RDD的DAG的形成过程，是通过依赖来完成的，每一个RDD通过转换算子的时候都会生成一个和多个子RDD，在通过转换算子的时候，在创建一个新的RDD的时候，也会创建他们之间的依赖关系。因此他们是通过Dependencies连接起来的，RDD的依赖不是我们的重点，如果想了解RDD的依赖，可以自行google，RDD的依赖分为：1:1的OneToOneDependency，m:1的RangeDependency，还有m:n的ShuffleDependencies，其中OneToOneDependency和RangeDependency又被称为NarrowDependency，这里的1:1,m:1,m:n的粒度是对于RDD的分区而言的。
依赖中最根本的是保留了父RDD，其rdd的方法就是返回父RDD的方法。这样其就形成了一个链表形式的结构，通过最后面的RDD根据依赖，可以向前回溯到所有的父类RDD。
我们以map为例，来看一下依赖是如何产生的。
通过map其实其实创建了一个MapPartitonsRDD的RDD
然后我们看一下MapPartitonsRDD的主构造函数，其又对RDD进行了赋值，其中父RDD就是上面的this对象指定的RDD，我们再看一下RDD这个类的构造函数：
其又调用了RDD的主构造函数
其实依赖都是在RDD的构造函数中形成的。
通过上面的依赖转换就形成了RDD额DAG图
生成了一个RDD的DAG图：
spark的job的划分、stage的划分
spark的Application划分job其实挺简单的，一个Application划分为几个job，我们就要看这个Application中有多少个Action算子，一个Action算子对应一个job，这个可以通过源码来看出来，转换算子是形成一个或者多个RDD，而Action算子是触发job的提交。
比如上面的map转换算子就是这样的
而Action算子是这样的：
通过runJob方法提交作业。stage的划分是根据是否进行shuflle过程来决定的，这个后面会细说。
spark的DAGScheduler的调度
当我们通过客户端，向spark集群提交作业时，如果利用的资源管理器是yarn，那么客户端向spark提交申请运行driver进程的机器，driver其实在spark中是没有具体的类的，driver机器主要是用来运行用户编写的代码的地方，完成DAGScheduler和TaskSchedule，追踪task运行的状态。记住，用户编写的主函数是在driver中运行的，但是RDD转换和执行是在不同的机器上完成。其实driver主要负责作业的调度和分发。Action算子到stage的划分和DAGScheduler的完成过程。
当我们在driver进程中运行用户定义的main函数的时候，首先会创建SparkContext对象，这个是我们与spark集群进行交互的入口它会初始化很多运行需要的环境，最主要的是初始化了DAGScheduler和TaskSchedule。
我们以这样的的一个RDD的逻辑执行图来分析整个DAGScheduler的过程。
因为DAGScheduler发生在driver进程中，我们就冲Driver进程运行用户定义的main函数开始。在上图中RDD9是最后一个RDD并且其调用了Action算子，就会触发作业的提交，其会调用SparkContext的runjob函数，其经过一系列的runJob的封装，会调用DAGScheduler的runJob
在SparkContext中存在着runJob方法
———————————————-
def runJob[T, U: ClassTag](
rdd: RDD[T], // rdd为上面提到的RDD逻辑执行图中的RDD9
func: (TaskContext, Iterator[T]) => U,这个方法也是RDD9调用Action算子传入的函数
partitions: Seq[Int],
resultHandler: (Int, U) => Unit): Unit = {
if (stopped.get()) {
throw new IllegalStateException("SparkContext has been shutdown")
}
val callSite = getCallSite
val cleanedFunc = clean(func)
logInfo("Starting job: " + callSite.shortForm)
if (conf.getBoolean("spark.logLineage", false)) {
logInfo("RDD’s recursive dependencies:/n" + rdd.toDebugString)
}
dagScheduler.runJob(rdd, cleanedFunc, partitions, callSite, resultHandler, localProperties.get)
progressBar.foreach(_.finishAll())
rdd.doCheckpoint()
}
———————————————-
DAGScheduler的runJob
———————————————-
def runJob[T, U](
rdd: RDD[T], //RDD9
func: (TaskContext, Iterator[T]) => U,
partitions: Seq[Int],
callSite: CallSite,
resultHandler: (Int, U) => Unit,
properties: Properties): Unit = {
val start = System.nanoTime
//在这里会生成一个job的守护进程waiter，用来等待作业提交执行是否完成，其又调用了submitJob，其以下的代
//码都是用来处运行结果的一些log日志信息
val waiter = submitJob(rdd, func, partitions, callSite, resultHandler, properties)
ThreadUtils.awaitReady(waiter.completionFuture, Duration.Inf)
waiter.completionFuture.value.get match {
case scala.util.Success(_) =>
logInfo("Job %d finished: %s, took %f s".format
(waiter.jobId, callSite.shortForm, (System.nanoTime – start) / 1e9))
case scala.util.Failure(exception) =>
logInfo("Job %d failed: %s, took %f s".format
(waiter.jobId, callSite.shortForm, (System.nanoTime – start) / 1e9))
// SPARK-8644: Include user stack trace in exceptions coming from DAGScheduler.
val callerStackTrace = Thread.currentThread().getStackTrace.tail
exception.setStackTrace(exception.getStackTrace ++ callerStackTrace)
throw exception
}
}
———————————————-
submitJob的源代码
———————————————-
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] = {
// 检查RDD的分区是否合法
val maxPartitions = rdd.partitions.length
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)
}
val jobId = nextJobId.getAndIncrement()
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)
//这一块是把我们的job继续进行封装到JobSubmitted，然后放入到一个进程中池里，spark会启动一个线程来处理我
//们提交的作业
val func2 = func.asInstanceOf[(TaskContext, Iterator[]) => ]
val waiter = new JobWaiter(this, jobId, partitions.size, resultHandler)
eventProcessLoop.post(JobSubmitted(
jobId, rdd, func2, partitions.toArray, callSite, waiter,
SerializationUtils.clone(properties)))
waiter
}
———————————————-
在DAGScheduler类中有一个DAGSchedulerEventProcessLoop的类，用来接收处理DAGScheduler的消息事件
JobSubmitted对象，因此会执行第一个操作handleJobSubmitted，在这里我们要说一下，Stage的类型，在spark中有两种类型的stage一种是ShuffleMapStage，和ResultStage，最后一个RDD对应的Stage是ResultStage，遇到Shuffle过程的RDD被称为ShuffleMapStage。
———————————————-
private[scheduler] def handleJobSubmitted(jobId: Int,
finalRDD: RDD[],//对应RDD9
func: (TaskContext, Iterator[]) => _,
partitions: Array[Int],
callSite: CallSite,
listener: JobListener,
properties: Properties) {
var finalStage: ResultStage = null
try {
// 先创建ResultStage。
finalStage = createResultStage(finalRDD, func, partitions, jobId, callSite)
} catch {
case e: Exception =>
logWarning("Creating new stage failed due to exception – job: " + jobId, e)
listener.jobFailed(e)
return
}
val job = new ActiveJob(jobId, finalStage, callSite, listener, properties)
clearCacheLocs()
logInfo("Got job %s (%s) with %d output partitions".format(
job.jobId, callSite.shortForm, partitions.length))
logInfo("Final stage: " + finalStage + " (" + finalStage.name + ")")
logInfo("Parents of final stage: " + finalStage.parents)
logInfo("Missing parents: " + getMissingParentStages(finalStage))
val jobSubmissionTime = clock.getTimeMillis()
jobIdToActiveJob(jobId) = job
activeJobs += job
finalStage.setActiveJob(job)
val stageIds = jobIdToStageIds(jobId).toArray
val stageInfos = stageIds.flatMap(id => stageIdToStage.get(id).map(_.latestInfo))
listenerBus.post(
SparkListenerJobStart(job.jobId, jobSubmissionTime, stageInfos, properties))
submitStage(finalStage)
}
———————————————-
上面的createResultStage其实就是RDD转换为Stage的过程，方法如下
———————————————-
/*
创建ResultStage的时候，它会调用相关函数
*/
private def createResultStage(
rdd: RDD[], //对应上图的RDD9
func: (TaskContext, Iterator[]) => _,
partitions: Array[Int],
jobId: Int,
callSite: CallSite): ResultStage = {
val parents = getOrCreateParentStages(rdd, jobId)
val id = nextStageId.getAndIncrement()
val stage = new ResultStage(id, rdd, func, partitions, parents, jobId, callSite)
stageIdToStage(id) = stage
updateJobIdStageIdMaps(jobId, stage)
stage
}
/**
- 形成ResultStage依赖的父Stage
*/
private def getOrCreateParentStages(rdd: RDD[_], firstJobId: Int): List[Stage] = {
getShuffleDependencies(rdd).map { shuffleDep =>
getOrCreateShuffleMapStage(shuffleDep, firstJobId)
}.toList
}
/** - 采用的是深度优先遍历找到Action算子的父依赖中的宽依赖
- 这个是最主要的方法，要看懂这个方法，其实后面的就好理解，最好结合这例子上面给出的RDD逻辑依赖图，比*
-
较容易看出来，根据上面的RDD逻辑依赖图，其返回的ShuffleDependency就是RDD2和RDD1，RDD7和RDD6的依
赖,如果存在A<-B<-C,这两个都是shuffle依赖，那么对于C其只返回B的shuffle依赖，而不会返回A
/
private[scheduler] def getShuffleDependencies(
rdd: RDD[]): HashSet[ShuffleDependency[, , ]] = {
//用来存放依赖
val parents = new HashSet[ShuffleDependency[, , ]]
//遍历过的RDD放入这个里面
val visited = new HashSet[RDD[]]
//创建一个待遍历RDD的栈结构
val waitingForVisit = new ArrayStack[RDD[]]
//压入finalRDD，逻辑图中的RDD9
waitingForVisit.push(rdd)
//循环遍历这个栈结构
while (waitingForVisit.nonEmpty) {
val toVisit = waitingForVisit.pop()
// 如果RDD没有被遍历过执行其中的代码
if (!visited(toVisit)) {
//然后把其放入已经遍历队列中
visited += toVisit
//得到依赖，我们知道依赖中存放的有父RDD的对象
toVisit.dependencies.foreach {
//如果这个依赖是shuffle依赖，则放入返回队列中
case shuffleDep: ShuffleDependency[, , ] =>
parents += shuffleDep
case dependency =>
//如果不是shuffle依赖，把其父RDD压入待访问栈中，从而进行循环
waitingForVisit.push(dependency.rdd)
}
}
}
parents
}
/创建shuffleMapStage，根据上面得到的两个Shuffle对象，分别创建了两个shuffleMapStage
/
/
def createShuffleMapStage(shuffleDep: ShuffleDependency[, , _], jobId: Int): ShuffleMapStage = {
//这个RDD其实就是RDD1和RDD6
val rdd = shuffleDep.rdd
val numTasks = rdd.partitions.length
val parents = getOrCreateParentStages(rdd, jobId) //查看这两个ShuffleMapStage是否存在父Shuffle的Stage
val id = nextStageId.getAndIncrement()
//创建ShuffleMapStage，下面是更新一下SparkContext的状态
val stage = new ShuffleMapStage(
id, rdd, numTasks, parents, jobId, rdd.creationSite, shuffleDep, mapOutputTracker)
stageIdToStage(id) = stage
shuffleIdToMapStage(shuffleDep.shuffleId) = stage
updateJobIdStageIdMaps(jobId, stage)if (!mapOutputTracker.containsShuffle(shuffleDep.shuffleId)) {
// Kind of ugly: need to register RDDs with the cache and map output tracker here
// since we can’t do it in the RDD constructor because # of partitions is unknown
logInfo("Registering RDD " + rdd.id + " (" + rdd.getCreationSite + ")")
mapOutputTracker.registerShuffle(shuffleDep.shuffleId, rdd.partitions.length)
}
stage
}———————————————-
通过上面的源代码分析，结合RDD的逻辑执行图，我们可以看出，这个job拥有三个Stage，一个ResultStage，两个ShuffleMapStage，一个ShuffleMapStage中的RDD是RDD1，另一个stage中的RDD是RDD6,从而，以上完成了RDD到Stage的切分工作。当切分完成后在handleJobSubmitted这个方法的最后，调用提交stage的方法。
submitStage源代码比较简单，它会检查我们当前的stage依赖的父stage是否已经执行完成，如果没有执行完成会循环提交其父stage等待其父stage执行完成了，才提交我们当前的stage进行执行。
———————————————-
private def submitStage(stage: Stage) {
val jobId = activeJobForStage(stage)
if (jobId.isDefined) {
logDebug("submitStage(" + stage + ")")
if (!waitingStages(stage) && !runningStages(stage) && !failedStages(stage)) {
val missing = getMissingParentStages(stage).sortBy(_.id)
logDebug("missing: " + missing)
if (missing.isEmpty) {
logInfo("Submitting " + stage + " (" + stage.rdd + "), which has no missing parents")
submitMissingTasks(stage, jobId.get)
} else {
for (parent <- missing) {
submitStage(parent)
}
waitingStages += stage
}
}
} else {
abortStage(stage, "No active job for stage " + stage.id, None)
}
}
———————————————-
提交task的方法源代码，我们按照刚才的三个stage中，提交的是前两个stage的过程来看待这个源代码。以包含RDD1的stage为例
———————————————-
private def submitMissingTasks(stage: Stage, jobId: Int) {
logDebug("submitMissingTasks(" + stage + ")")
// Get our pending tasks and remember them in our pendingTasks entry
stage.pendingPartitions.clear()
// 计算需要计算的分区数
val partitionsToCompute: Seq[Int] = stage.findMissingPartitions()
// Use the scheduling pool, job group, description, etc. from an ActiveJob associated
// with this Stage
val properties = jobIdToActiveJob(jobId).properties
runningStages += stage
// 封装stage的一些信息，得到stage到分区数的映射关系，即一个stage对应多少个分区需要计算
stage match {
case s: ShuffleMapStage =>
outputCommitCoordinator.stageStart(stage = s.id, maxPartitionId = s.numPartitions - 1)
case s: ResultStage =>
outputCommitCoordinator.stageStart(
stage = s.id, maxPartitionId = s.rdd.partitions.length - 1)
}//得到每个分区对应的具体位置，即分区的数据位于集群的哪台机器上。
val taskIdToLocations: Map[Int, Seq[TaskLocation]] = try {
stage match {
case s: ShuffleMapStage =>
partitionsToCompute.map { id => (id, getPreferredLocs(stage.rdd, id))}.toMap
case s: ResultStage =>
partitionsToCompute.map { id =>
val p = s.partitions(id)
(id, getPreferredLocs(stage.rdd, p))
}.toMap
}
} catch {
case NonFatal(e) =>
stage.makeNewStageAttempt(partitionsToCompute.size)
listenerBus.post(SparkListenerStageSubmitted(stage.latestInfo, properties))
abortStage(stage, s"Task creation failed: $e/n${Utils.exceptionString(e)}", Some(e))
runningStages -= stage
return
}
// 这个把上面stage要计算的分区和每个分区对应的物理位置进行了从新封装，放在了latestInfo里面
stage.makeNewStageAttempt(partitionsToCompute.size, taskIdToLocations.values.toSeq)
listenerBus.post(SparkListenerStageSubmitted(stage.latestInfo, properties))
//序列化我们刚才得到的信息，以便在driver机器和work机器之间进行传输
var taskBinary: Broadcast[Array[Byte]] = null
try {
// For ShuffleMapTask, serialize and broadcast (rdd, shuffleDep).
// For ResultTask, serialize and broadcast (rdd, func).
val taskBinaryBytes: Array[Byte] = stage match {
case stage: ShuffleMapStage =>
JavaUtils.bufferToArray(
closureSerializer.serialize((stage.rdd, stage.shuffleDep): AnyRef))
case stage: ResultStage =>
JavaUtils.bufferToArray(closureSerializer.serialize((stage.rdd, stage.func): AnyRef))
}
taskBinary = sc.broadcast(taskBinaryBytes)
} catch {
// In the case of a failure during serialization, abort the stage.
case e: NotSerializableException =>
abortStage(stage, "Task not serializable: " + e.toString, Some(e))
runningStages -= stage
// Abort execution
return
case NonFatal(e) =>
abortStage(stage, s"Task serialization failed: $e/n${Utils.exceptionString(e)}", Some(e))
runningStages -= stage
return
}//封装stage构成taskSet集合，ShuffleMapStage对应的task为ShuffleMapTask，而ResultStage对应的taskSet为ResultTask
val tasks: Seq[Task[_]] = try {
stage match {
case stage: ShuffleMapStage =>
partitionsToCompute.map { id =>
val locs = taskIdToLocations(id)
val part = stage.rdd.partitions(id)
new ShuffleMapTask(stage.id, stage.latestInfo.attemptId,
taskBinary, part, locs, stage.latestInfo.taskMetrics, properties, Option(jobId),
Option(sc.applicationId), sc.applicationAttemptId)
}
case stage: ResultStage =>
partitionsToCompute.map { id =>
val p: Int = stage.partitions(id)
val part = stage.rdd.partitions(p)
val locs = taskIdToLocations(id)
new ResultTask(stage.id, stage.latestInfo.attemptId,
taskBinary, part, locs, id, properties, stage.latestInfo.taskMetrics,
Option(jobId), Option(sc.applicationId), sc.applicationAttemptId)
}
}} catch {
case NonFatal(e) =>
abortStage(stage, s"Task creation failed: $e/n${Utils.exceptionString(e)}", Some(e))
runningStages -= stage
return
}
//提交task给TaskSchedule
if (tasks.size > 0) {
logInfo("Submitting " + tasks.size + " missing tasks from " + stage + " (" + stage.rdd + ")")
stage.pendingPartitions ++= tasks.map(_.partitionId)
logDebug("New pending partitions: " + stage.pendingPartitions)
taskScheduler.submitTasks(new TaskSet(
tasks.toArray, stage.id, stage.latestInfo.attemptId, jobId, properties))
stage.latestInfo.submissionTime = Some(clock.getTimeMillis())
} else {
// Because we posted SparkListenerStageSubmitted earlier, we should mark
// the stage as completed here in case there are no tasks to run
markStageAsFinished(stage, None)
val debugString = stage match {
case stage: ShuffleMapStage =>
s"Stage ${stage} is actually done; " +
s"(available: ${stage.isAvailable}," +
s"available outputs: ${stage.numAvailableOutputs}," +
s"partitions: ${stage.numPartitions})"
case stage : ResultStage =>
s"Stage ${stage} is actually done; (partitions: ${stage.numPartitions})"
}
logDebug(debugString)
submitWaitingChildStages(stage)}
}
———————————————-
到此，完成了整个DAGScheduler的调度。
spark的TaskSchedule的调度
spark的Task的调度，我们要明白其调度过程，其根据不同的资源管理器拥有不同的调度策略，因此也拥有不同的调度守护进程，这个守护进程管理着集群的资源信息，spark提供了一个基本的守护进程的类，来完成与driver和executor的交互：CoarseGrainedSchedulerBackend，它应该运行在集群资源管理器上，比如yarn等。他收集了集群work机器的一般资源信息。当我们形成tasks将要进行调度的时候，driver进程会与其通信，请求资源的分配和调度，其会把最优的work节点分配给task来执行其任务。而TaskScheduleImpl实现了task调度的过程，采用的调度算法默认的是FIFO的策略，也可以采用公平调度策略。
当我们提交task时，其会创建一个管理task的类TaskSetManager，然后把其加入到任务调度池中。
———————————————-
override def submitTasks(taskSet: TaskSet) {
val tasks = taskSet.tasks
logInfo("Adding task set " + taskSet.id + " with " + tasks.length + " tasks")
this.synchronized {
// 创建taskSetManager，以下为更新一下状态
val manager = createTaskSetManager(taskSet, maxTaskFailures)
val stage = taskSet.stageId
val stageTaskSets =
taskSetsByStageIdAndAttempt.getOrElseUpdate(stage, new HashMap[Int, TaskSetManager])
stageTaskSets(taskSet.stageAttemptId) = manager
val conflictingTaskSet = stageTaskSets.exists { case (, ts) =>
ts.taskSet != taskSet && !ts.isZombie
}
if (conflictingTaskSet) {
throw new IllegalStateException(s"more than one active taskSet for stage $stage:" +
s" ${stageTaskSets.toSeq.map{._2.taskSet.id}.mkString(",")}")
}
//把封装好的taskSet，加入到任务调度队列中。
schedulableBuilder.addTaskSetManager(manager, manager.taskSet.properties)
if (!isLocal && !hasReceivedTask) {
starvationTimer.scheduleAtFixedRate(new TimerTask() {
override def run() {
if (!hasLaunchedTask) {
logWarning("Initial job has not accepted any resources; " +
"check your cluster UI to ensure that workers are registered " +
"and have sufficient resources")
} else {
this.cancel()
}
}
}, STARVATION_TIMEOUT_MS, STARVATION_TIMEOUT_MS)
}
hasReceivedTask = true}
//这个地方就是向资源管理器发出请求，请求任务的调度
backend.reviveOffers()
}
/*
*这个方法是位于CoarseGrainedSchedulerBackend类中，driver进程会想集群管理器发送请求资源的请求。
/
override def reviveOffers() {
driverEndpoint.send(ReviveOffers)
}
———————————————-
当其收到这个请求时，其会调用这样的方法。
———————————————-
override def receive: PartialFunction[Any, Unit] = {
case StatusUpdate(executorId, taskId, state, data) =>
scheduler.statusUpdate(taskId, state, data.value)
if (TaskState.isFinished(state)) {
executorDataMap.get(executorId) match {
case Some(executorInfo) =>
executorInfo.freeCores += scheduler.CPUS_PER_TASK
makeOffers(executorId)
case None =>
// Ignoring the update since we don’t know about the executor.
logWarning(s"Ignored task status update ($taskId state $state) " +
s"from unknown executor with ID $executorId")
}
}
//发送的请求满足这个条件
case ReviveOffers =>
makeOffers()
case KillTask(taskId, executorId, interruptThread) =>
executorDataMap.get(executorId) match {
case Some(executorInfo) =>
executorInfo.executorEndpoint.send(KillTask(taskId, executorId, interruptThread))
case None =>
// Ignoring the task kill since the executor is not registered.
logWarning(s"Attempted to kill task $taskId for unknown executor $executorId.")
}
}
/*
*这个方法是搜集集群上现在还在活着的机器的相关信息。并且进行封装成WorkerOffer类，
- 然后其会调用TaskSchedulerImpl中的resourceOffers方法，来进行筛选，筛选出符合请求资源的机器，来执行我们当前的任务
/
private def makeOffers() {
// Filter out executors under killing
val activeExecutors = executorDataMap.filterKeys(executorIsAlive)
val workOffers = activeExecutors.map { case (id, executorData) =>
new WorkerOffer(id, executorData.executorHost, executorData.freeCores)
}.toIndexedSeq
launchTasks(scheduler.resourceOffers(workOffers))
}
/*
得到集群中空闲机器的信息后，我们通过此方法来筛选出满足我们这次任务要求的机器，然后返回TaskDescription类
*这个类封装了task与excutor的相关信息
-
/
def resourceOffers(offers: IndexedSeq[WorkerOffer]): Seq[Seq[TaskDescription]] = synchronized {
// Mark each slave as alive and remember its hostname
// Also track if new executor is added
var newExecAvail = false
//检查work是否已经存在了，把不存在的加入到work调度池中
for (o <- offers) {
if (!hostToExecutors.contains(o.host)) {
hostToExecutors(o.host) = new HashSet[String]()
}
if (!executorIdToRunningTaskIds.contains(o.executorId)) {
hostToExecutors(o.host) += o.executorId
executorAdded(o.executorId, o.host)
executorIdToHost(o.executorId) = o.host
executorIdToRunningTaskIds(o.executorId) = HashSet[Long]()
newExecAvail = true
}
for (rack <- getRackForHost(o.host)) {
hostsByRack.getOrElseUpdate(rack, new HashSet[String]()) += o.host
}
}
// 打乱work机器的顺序，以免每次分配任务时都在同一个机器上进行。避免某一个work计算压力太大。
val shuffledOffers = Random.shuffle(offers)
//对于每一work，创建一个与其核数大小相同的数组，数组的大小决定了这台work上可以并行执行task的数目.
val tasks = shuffledOffers.map(o => new ArrayBufferTaskDescription)
//取出每台机器的cpu核数
val availableCpus = shuffledOffers.map(o => o.cores).toArray
//从task任务调度池中，按照我们的调度算法，取出需要执行的任务
val sortedTaskSets = rootPool.getSortedTaskSetQueue
for (taskSet <- sortedTaskSets) {
logDebug("parentName: %s, name: %s, runningTasks: %s".format(
taskSet.parent.name, taskSet.name, taskSet.runningTasks))
if (newExecAvail) {
taskSet.executorAdded()
}
}
// 下面的这个循环，是用来标记task根据work的信息来标定数据本地化的程度的。当我们在yarn资源管理器，以–driver-mode配置
//为client时，我们就会在打出来的日志上看出每一台机器上运行task的数据本地化程度。同时还会选择每个task对应的work机器
// NOTE: the preferredLocality order: PROCESS_LOCAL, NODE_LOCAL, NO_PREF, RACK_LOCAL, ANY
for (taskSet <- sortedTaskSets) {
var launchedAnyTask = false
var launchedTaskAtCurrentMaxLocality = false
for (currentMaxLocality <- taskSet.myLocalityLevels) {
do {
launchedTaskAtCurrentMaxLocality = resourceOfferSingleTaskSet(
taskSet, currentMaxLocality, shuffledOffers, availableCpus, tasks)
launchedAnyTask |= launchedTaskAtCurrentMaxLocality
} while (launchedTaskAtCurrentMaxLocality)
}
if (!launchedAnyTask) {
taskSet.abortIfCompletelyBlacklisted(hostToExecutors)
}
}if (tasks.size > 0) {
hasLaunchedTask = true
}
//返回taskDescription对象
return tasks
}
/*
task选择执行其任务的work其实是在这个函数中实现的，从这个可以看出，一台work上其实是可以运行多个task，主要是看如何
*进行算法调度
-
/
private def resourceOfferSingleTaskSet(
taskSet: TaskSetManager,
maxLocality: TaskLocality,
shuffledOffers: Seq[WorkerOffer],
availableCpus: Array[Int],
tasks: IndexedSeq[ArrayBuffer[TaskDescription]]) : Boolean = {
var launchedTask = false
//循环所有的机器，找适合此机器的task
for (i <- 0 until shuffledOffers.size) {
val execId = shuffledOffers(i).executorId
val host = shuffledOffers(i).host
//判断其剩余的cpu核数是否满足我们的最低配置，满足则为其分配任务，否则不为其分配任务。
if (availableCpus(i) >= CPUS_PER_TASK) {
try {
//这个for中的resourOffer就是来判断其标记任务数据本地化的程度的。task(i)其实是一个数组，数组大小和其cpu核心数大小相同。
for (task <- taskSet.resourceOffer(execId, host, maxLocality)) {
tasks(i) += task
val tid = task.taskId
taskIdToTaskSetManager(tid) = taskSet
taskIdToExecutorId(tid) = execId
executorIdToRunningTaskIds(execId).add(tid)
availableCpus(i) -= CPUS_PER_TASK
assert(availableCpus(i) >= 0)
launchedTask = true
}
} catch {
case e: TaskNotSerializableException =>
logError(s"Resource offer failed, task set ${taskSet.name} was not serializable")
// Do not offer resources for this task, but don’t throw an error to allow other
// task sets to be submitted.
return launchedTask
}
}
}
return launchedTask
}———————————————-
以上完成了从TaskSet到task和work机器的绑定过程的所有任务。下面就是如何发送task到executor进行执行。在makeOffers()方法中调用了launchTasks方法,这个方法其实就是发送task作业到指定的机器上。只此，spark TaskSchedule的调度就此结束。
executor如何执行task以及我们定义的函数
当TaskSchedule完成对task的调度时，task需要在work机器上来进行执行。此时，work机器就会启动一个Backend的守护进程，用来完成与driver和资源管理器的通信。这个Backend就是CoarseGrainedExecutorBackend，启动的main主函数为,从main函数中可以看出，其主要进行参数的解析，然后运行run方法。
———————————————-
def main(args: Array[String]) {
var driverUrl: String = null
var executorId: String = null
var hostname: String = null
var cores: Int = 0
var appId: String = null
var workerUrl: Option[String] = None
val userClassPath = new mutable.ListBuffer[URL]()
var argv = args.toList
while (!argv.isEmpty) {
argv match {
case ("–driver-url") :: value :: tail =>
driverUrl = value
argv = tail
case ("–executor-id") :: value :: tail =>
executorId = value
argv = tail
case ("–hostname") :: value :: tail =>
hostname = value
argv = tail
case ("–cores") :: value :: tail =>
cores = value.toInt
argv = tail
case ("–app-id") :: value :: tail =>
appId = value
argv = tail
case ("–worker-url") :: value :: tail =>
// Worker url is used in spark standalone mode to enforce fate-sharing with worker
workerUrl = Some(value)
argv = tail
case ("–user-class-path") :: value :: tail =>
userClassPath += new URL(value)
argv = tail
case Nil =>
case tail =>
// scalastyle:off println
System.err.println(s"Unrecognized options: ${tail.mkString(" ")}")
// scalastyle:on println
printUsageAndExit()
}
}
if (driverUrl == null || executorId == null || hostname == null || cores <= 0 ||
appId == null) {
printUsageAndExit()
}
run(driverUrl, executorId, hostname, cores, appId, workerUrl, userClassPath)
System.exit(0)
}
/*
可以看出，run方法只是进行了一些需要运行task所需要的环境进行配置。并且创建相应的运行环境。
-
/
private def run(
driverUrl: String,
executorId: String,
hostname: String,
cores: Int,
appId: String,
workerUrl: Option[String],
userClassPath: Seq[URL]) {Utils.initDaemon(log)
SparkHadoopUtil.get.runAsSparkUser { () =>
// Debug code
Utils.checkHost(hostname)// Bootstrap to fetch the driver’s Spark properties.
val executorConf = new SparkConf
val port = executorConf.getInt("spark.executor.port", 0)
val fetcher = RpcEnv.create(
"driverPropsFetcher",
hostname,
port,
executorConf,
new SecurityManager(executorConf),
clientMode = true)
val driver = fetcher.setupEndpointRefByURI(driverUrl)
val cfg = driver.askWithRetrySparkAppConfig
val props = cfg.sparkProperties ++ Seq[(String, String)](("spark.app.id", appId))
fetcher.shutdown()// Create SparkEnv using properties we fetched from the driver.
val driverConf = new SparkConf()
for ((key, value) <- props) {
// this is required for SSL in standalone mode
if (SparkConf.isExecutorStartupConf(key)) {
driverConf.setIfMissing(key, value)
} else {
driverConf.set(key, value)
}
}
if (driverConf.contains("spark.yarn.credentials.file")) {
logInfo("Will periodically update credentials from: " +
driverConf.get("spark.yarn.credentials.file"))
SparkHadoopUtil.get.startCredentialUpdater(driverConf)
}val env = SparkEnv.createExecutorEnv(
driverConf, executorId, hostname, port, cores, cfg.ioEncryptionKey, isLocal = false)env.rpcEnv.setupEndpoint("Executor", new CoarseGrainedExecutorBackend(
env.rpcEnv, driverUrl, executorId, hostname, cores, userClassPath, env))
workerUrl.foreach { url =>
env.rpcEnv.setupEndpoint("WorkerWatcher", new WorkerWatcher(env.rpcEnv, url))
}
env.rpcEnv.awaitTermination()
SparkHadoopUtil.get.stopCredentialUpdater()
}
}———————————————-
其执行函数的调用过程如下：
我们知道当我们完成TaskSchedule的调度时，是通过rpc发送了一个消息，如下图所示，当work机器的Backend启动以后，其会与driver进程进行rpc通信，当其收到LaunchTask的消息后，其会执行下面的代码。
我们可以看出此方法存在很多的情况，根据接收到的不同的消息，执行不同的代码。我们上面执行的是LaunchTask的请求。
———————————————-
override def receive: PartialFunction[Any, Unit] = {
case RegisteredExecutor =>
logInfo("Successfully registered with driver")
try {
executor = new Executor(executorId, hostname, env, userClassPath, isLocal = false)
} catch {
case NonFatal(e) =>
exitExecutor(1, "Unable to create executor due to " + e.getMessage, e)
}
case RegisterExecutorFailed(message) =>
exitExecutor(1, "Slave registration failed: " + message)
//提交任务时，执行这样的操作。
case LaunchTask(data) =>
if (executor == null) {
exitExecutor(1, "Received LaunchTask command but executor was null")
} else {
//先反序列化
val taskDesc = ser.deserializeTaskDescription
logInfo("Got assigned task " + taskDesc.taskId)
//然后执行launchTask操作。
executor.launchTask(this, taskId = taskDesc.taskId, attemptNumber = taskDesc.attemptNumber,
taskDesc.name, taskDesc.serializedTask)
}
case KillTask(taskId, _, interruptThread) =>
if (executor == null) {
exitExecutor(1, "Received KillTask command but executor was null")
} else {
executor.killTask(taskId, interruptThread)
}
case StopExecutor =>
stopping.set(true)
logInfo("Driver commanded a shutdown")
// Cannot shutdown here because an ack may need to be sent back to the caller. So send
// a message to self to actually do the shutdown.
self.send(Shutdown)
case Shutdown =>
stopping.set(true)
new Thread("CoarseGrainedExecutorBackend-stop-executor") {
override def run(): Unit = {
// executor.stop() will call SparkEnv.stop() which waits until RpcEnv stops totally.
// However, if executor.stop() runs in some thread of RpcEnv, RpcEnv won’t be able to
// stop until executor.stop() returns, which becomes a dead-lock (See SPARK-14180).
// Therefore, we put this line in a new thread.
executor.stop()
}
}.start()
}
———————————————-
Executor的相关源代码,从源码中我们可以看出，对于Task，其创建了一个TaskRunner的线程，并且把其放入到执行队列中进行执行。
———————————————-
def launchTask(
context: ExecutorBackend,
taskId: Long,
attemptNumber: Int,
taskName: String,
serializedTask: ByteBuffer): Unit = {
val tr = new TaskRunner(context, taskId = taskId, attemptNumber = attemptNumber, taskName,
serializedTask)
runningTasks.put(taskId, tr)
threadPool.execute(tr)
}
———————————————-
从下面可以看出，其定义的就是一个线程，那我们就看一下这个线程的run方法。
———————————————-
override def run(): Unit = {
//初始化线程运行需要的一些环境
val threadMXBean = ManagementFactory.getThreadMXBean
val taskMemoryManager = new TaskMemoryManager(env.memoryManager, taskId)
val deserializeStartTime = System.currentTimeMillis()
val deserializeStartCpuTime = if (threadMXBean.isCurrentThreadCpuTimeSupported) {
threadMXBean.getCurrentThreadCpuTime
} else 0L
//得到当前进程的类加载器
Thread.currentThread.setContextClassLoader(replClassLoader)
val ser = env.closureSerializer.newInstance()
logInfo(s"Running $taskName (TID $taskId)")
//更新相关的状态
execBackend.statusUpdate(taskId, TaskState.RUNNING, EMPTY_BYTE_BUFFER)
var taskStart: Long = 0
var taskStartCpu: Long = 0
startGCTime = computeTotalGcTime()
try {//反序列化类相关的依赖，得到相关的参数
val (taskFiles, taskJars, taskProps, taskBytes) =
Task.deserializeWithDependencies(serializedTask)
// Must be set before updateDependencies() is called, in case fetching dependencies
// requires access to properties contained within (e.g. for access control).
Executor.taskDeserializationProps.set(taskProps)//更新依赖配置
updateDependencies(taskFiles, taskJars)
task = ser.deserialize[Task[Any]](taskBytes, Thread.currentThread.getContextClassLoader)
task.localProperties = taskProps
task.setTaskMemoryManager(taskMemoryManager)
// If this task has been killed before we deserialized it, let's quit now. Otherwise,
// continue executing the task.
if (killed) {
// Throw an exception rather than returning, because returning within a try{} block
// causes a NonLocalReturnControl exception to be thrown. The NonLocalReturnControl
// exception will be caught by the catch block, leading to an incorrect ExceptionFailure
// for the task.
throw new TaskKilledException
}
logDebug("Task " + taskId + "'s epoch is " + task.epoch)//追踪缓存数据的位置
env.mapOutputTracker.updateEpoch(task.epoch)
// Run the actual task and measure its runtime.
taskStart = System.currentTimeMillis()
taskStartCpu = if (threadMXBean.isCurrentThreadCpuTimeSupported) {
threadMXBean.getCurrentThreadCpuTime
} else 0L
var threwException = true//运行任务的run方法来运行task，主要就是下面的task.run方法，它又会调用runTask方法来真正执行task，前面我们提到过，job变
//为stage有两种，ShuffleMapStage和ResultStage，那么其对应的也有两个Task：ShuffleMapTask和ResultTask，不同的task类型，执行不同的run方法。
val value = try {
val res = task.run(
taskAttemptId = taskId,
attemptNumber = attemptNumber,
metricsSystem = env.metricsSystem)
threwException = false
res
} finally {
//下面就是根据上面的运行结果，来进行一些判断和日志的打出
val releasedLocks = env.blockManager.releaseAllLocksForTask(taskId)
val freedMemory = taskMemoryManager.cleanUpAllAllocatedMemory()
if (freedMemory > 0 && !threwException) {
val errMsg = s"Managed memory leak detected; size = $freedMemory bytes, TID = $taskId"
if (conf.getBoolean("spark.unsafe.exceptionOnMemoryLeak", false)) {
throw new SparkException(errMsg)
} else {
logWarning(errMsg)
}
}
if (releasedLocks.nonEmpty && !threwException) {
val errMsg =
s"${releasedLocks.size} block locks were not released by TID = $taskId:/n" +
releasedLocks.mkString("[", ", ", "]")
if (conf.getBoolean("spark.storage.exceptionOnPinLeak", false)) {
throw new SparkException(errMsg)
} else {
logWarning(errMsg)
}
}
}
val taskFinish = System.currentTimeMillis()
val taskFinishCpu = if (threadMXBean.isCurrentThreadCpuTimeSupported) {
threadMXBean.getCurrentThreadCpuTime
} else 0L
// If the task has been killed, let's fail it.
if (task.killed) {
throw new TaskKilledException
}
val resultSer = env.serializer.newInstance()
val beforeSerialization = System.currentTimeMillis()
val valueBytes = resultSer.serialize(value)
val afterSerialization = System.currentTimeMillis()
// Deserialization happens in two parts: first, we deserialize a Task object, which
// includes the Partition. Second, Task.run() deserializes the RDD and function to be run.
task.metrics.setExecutorDeserializeTime(
(taskStart - deserializeStartTime) + task.executorDeserializeTime)
task.metrics.setExecutorDeserializeCpuTime(
(taskStartCpu - deserializeStartCpuTime) + task.executorDeserializeCpuTime)
// We need to subtract Task.run()'s deserialization time to avoid double-counting
task.metrics.setExecutorRunTime((taskFinish - taskStart) - task.executorDeserializeTime)
task.metrics.setExecutorCpuTime(
(taskFinishCpu - taskStartCpu) - task.executorDeserializeCpuTime)
task.metrics.setJvmGCTime(computeTotalGcTime() - startGCTime)
task.metrics.setResultSerializationTime(afterSerialization - beforeSerialization)
// Note: accumulator updates must be collected after TaskMetrics is updated
val accumUpdates = task.collectAccumulatorUpdates()
// TODO: do not serialize value twice
val directResult = new DirectTaskResult(valueBytes, accumUpdates)
val serializedDirectResult = ser.serialize(directResult)
val resultSize = serializedDirectResult.limit
// directSend = sending directly back to the driver
val serializedResult: ByteBuffer = {
if (maxResultSize > 0 && resultSize > maxResultSize) {
logWarning(s"Finished $taskName (TID $taskId). Result is larger than maxResultSize " +
s"(${Utils.bytesToString(resultSize)} > ${Utils.bytesToString(maxResultSize)}), " +
s"dropping it.")
ser.serialize(new IndirectTaskResult[Any](TaskResultBlockId(taskId), resultSize))
} else if (resultSize > maxDirectResultSize) {
val blockId = TaskResultBlockId(taskId)
env.blockManager.putBytes(
blockId,
new ChunkedByteBuffer(serializedDirectResult.duplicate()),
StorageLevel.MEMORY_AND_DISK_SER)
logInfo(
s"Finished $taskName (TID $taskId). $resultSize bytes result sent via BlockManager)")
ser.serialize(new IndirectTaskResult[Any](blockId, resultSize))
} else {
logInfo(s"Finished $taskName (TID $taskId). $resultSize bytes result sent to driver")
serializedDirectResult
}
}
execBackend.statusUpdate(taskId, TaskState.FINISHED, serializedResult)
} catch {
case ffe: FetchFailedException =>
val reason = ffe.toTaskFailedReason
setTaskFinishedAndClearInterruptStatus()
execBackend.statusUpdate(taskId, TaskState.FAILED, ser.serialize(reason))
case _: TaskKilledException =>
logInfo(s"Executor killed $taskName (TID $taskId)")
setTaskFinishedAndClearInterruptStatus()
execBackend.statusUpdate(taskId, TaskState.KILLED, ser.serialize(TaskKilled))
case _: InterruptedException if task.killed =>
logInfo(s"Executor interrupted and killed $taskName (TID $taskId)")
setTaskFinishedAndClearInterruptStatus()
execBackend.statusUpdate(taskId, TaskState.KILLED, ser.serialize(TaskKilled))
case CausedBy(cDE: CommitDeniedException) =>
val reason = cDE.toTaskFailedReason
setTaskFinishedAndClearInterruptStatus()
execBackend.statusUpdate(taskId, TaskState.FAILED, ser.serialize(reason))
case t: Throwable =>
// Attempt to exit cleanly by informing the driver of our failure.
// If anything goes wrong (or this was a fatal exception), we will delegate to
// the default uncaught exception handler, which will terminate the Executor.
logError(s"Exception in $taskName (TID $taskId)", t)
// Collect latest accumulator values to report back to the driver
val accums: Seq[AccumulatorV2[_, _]] =
if (task != null) {
task.metrics.setExecutorRunTime(System.currentTimeMillis() - taskStart)
task.metrics.setJvmGCTime(computeTotalGcTime() - startGCTime)
task.collectAccumulatorUpdates(taskFailed = true)
} else {
Seq.empty
}
val accUpdates = accums.map(acc => acc.toInfo(Some(acc.value), None))
val serializedTaskEndReason = {
try {
ser.serialize(new ExceptionFailure(t, accUpdates).withAccums(accums))
} catch {
case _: NotSerializableException =>
// t is not serializable so just send the stacktrace
ser.serialize(new ExceptionFailure(t, accUpdates, false).withAccums(accums))
}
}
setTaskFinishedAndClearInterruptStatus()
execBackend.statusUpdate(taskId, TaskState.FAILED, serializedTaskEndReason)
// Don't forcibly exit unless the exception was inherently fatal, to avoid
// stopping other tasks unnecessarily.
if (Utils.isFatalError(t)) {
SparkUncaughtExceptionHandler.uncaughtException(t)
}
} finally {
runningTasks.remove(taskId)
}}
}
———————————————-
前面我们提到过，job变为stage有两种，ShuffleMapStage和ResultStage，那么其对应的也有两个Task：ShuffleMapTask和
ResultTask，不同的task类型，执行不同的Task.runTask方法。Task.run方法中调用了runTask的方法，这个方法在上面两个Task类中都进行了重写。
ShuffleMapTask的runTask方法
———————————————-
override def runTask(context: TaskContext): MapStatus = {
// Deserialize the RDD using the broadcast variable.
//首先进行一些初始化操作
val threadMXBean = ManagementFactory.getThreadMXBean
val deserializeStartTime = System.currentTimeMillis()
val deserializeStartCpuTime = if (threadMXBean.isCurrentThreadCpuTimeSupported) {
threadMXBean.getCurrentThreadCpuTime
} else 0L
val ser = SparkEnv.get.closureSerializer.newInstance()
//反序列化，这里的rdd，其实是我们进行shuffle之前的最后一个rdd，这个我们在前面已经说到的。
val (rdd, dep) = ser.deserialize[(RDD[], ShuffleDependency[, , ])](
ByteBuffer.wrap(taskBinary.value), Thread.currentThread.getContextClassLoader)
_executorDeserializeTime = System.currentTimeMillis() – deserializeStartTime
executorDeserializeCpuTime = if (threadMXBean.isCurrentThreadCpuTimeSupported) {
threadMXBean.getCurrentThreadCpuTime – deserializeStartCpuTime
} else 0L
//下面就是把每一个shuffle之前的stage的最后一个rdd进行写入操作，但是没有看到task执行我们写的函数，也没有看到其调用compute函数以及rdd之间的管道执行也没有体现出来，往下看，会揭露这些问题的面纱。
var writer: ShuffleWriter[Any, Any] = null
try {
val manager = SparkEnv.get.shuffleManager
writer = manager.getWriter[Any, Any](dep.shuffleHandle, partitionId, context)
writer.write(rdd.iterator(partition, context).asInstanceOf[Iterator[ <: Product2[Any, Any]]])
writer.stop(success = true).get
} catch {
case e: Exception =>
try {
if (writer != null) {
writer.stop(success = false)
}
} catch {
case e: Exception =>
log.debug("Could not stop writer", e)
}
throw e
}
}
———————————————-
对于上面红色部分的问题，我们在这里进行详细的解释。RDD会根据依赖关系来形成一个有向无环图，通过最后一个RDD和其依赖，我们就可以反向查找其对应的所有父类。如果没有shuffle过程，那么其就会形成管道，形成管道的好处就是所有RDD的中间结果不需要进行存储，直接就把我们的定义的多个函数串连起来，从输入到输出中间结果不需要存储，节省了时间和空间。同时我们也知道RDD的中间结果可以持久化到内存或者硬盘上，spark对于这个是可以追踪到的。
通过上面的分析，我们可以看出，executor中
正是我们RDD往前回溯的开始。对于shuffle过程和ResultTask的runTask的执行过程以后会在慢慢跟进。
原创文章，作者：kepupublish，如若转载，请注明出处：https://blog.ytso.com/191988.html