Spark Streaming从Flume Poll数据案例实战和内幕源码解密

本节课分成二部分讲解:

一、Spark Streaming on Polling from Flume实战

二、Spark Streaming on Polling from Flume源码

第一部分:

推模式(Flume push SparkStreaming) VS 拉模式(SparkStreaming poll Flume)

采用推模式:推模式的理解就是Flume作为缓存,存有数据。监听对应端口,如果服务可以链接,就将数据push过去。(简单,耦合要低),缺点是SparkStreaming 程序没有启动的话,Flume端会报错,同时会导致Spark Streaming 程序来不及消费的情况。

采用拉模式:拉模式就是自己定义一个sink,SparkStreaming自己去channel里面取数据,根据自身条件去获取数据,稳定性好。

Flume poll 实战:

1.Flume poll 配置

进入http://spark.apache.org/docs/latest/streaming-flume-integration.html官网,下载

spark-streaming-flume-sink_2.10-1.6.0.jar、scala-library-2.10.5.jar、commons-lang3-3.3.2.jar三个包:

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第1张图片

 

将下载后的三个jar包放入Flume安装lib目录:

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第2张图片 

配置Flume conf环境参数:

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第3张图片

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第4张图片

编写业务代码:

public class SparkStreamingPollDataFromFlume {

          public static void main(String[] args) {

                    /*

                     * 第一步:配置SparkConf:

                     * 1,至少2条线程:因为Spark Streaming应用程序在运行的时候,至少有一条

                     * 线程用于不断的循环接收数据,并且至少有一条线程用于处理接受的数据(否则的话无法

                     * 有线程用于处理数据,随着时间的推移,内存和磁盘都会不堪重负);

                     * 2,对于集群而言,每个Executor一般肯定不止一个Thread,那对于处理Spark Streaming的

                     * 应用程序而言,每个Executor一般分配多少Core比较合适?根据我们过去的经验,5个左右的

                     * Core是最佳的(一个段子分配为奇数个Core表现最佳,例如3个、5个、7个Core等);

                     */

                    SparkConf conf = new SparkConf().setAppName("SparkStreamingPollDataFromFlume").setMaster("local[2]");

                    /*

                     * 第二步:创建SparkStreamingContext:

                     * 1,这个是SparkStreaming应用程序所有功能的起始点和程序调度的核心

                     * SparkStreamingContext的构建可以基于SparkConf参数,也可基于持久化的SparkStreamingContext的内容

                     * 来恢复过来(典型的场景是Driver崩溃后重新启动,由于Spark Streaming具有连续7*24小时不间断运行的特征,

                     * 所有需要在Driver重新启动后继续上衣系的状态,此时的状态恢复需要基于曾经的Checkpoint);

                     * 2,在一个Spark Streaming应用程序中可以创建若干个SparkStreamingContext对象,使用下一个SparkStreamingContext

                     * 之前需要把前面正在运行的SparkStreamingContext对象关闭掉,由此,我们获得一个重大的启发SparkStreaming框架也只是

                     * Spark Core上的一个应用程序而已,只不过Spark Streaming框架箱运行的话需要Spark工程师写业务逻辑处理代码;

                     */

                    JavaStreamingContext jsc = new JavaStreamingContext(conf, Durations.seconds(30));

                    /*

                     * 第三步:创建Spark Streaming输入数据来源input Stream:

                     * 1,数据输入来源可以基于File、HDFS、Flume、Kafka、Socket等

                     * 2, 在这里我们指定数据来源于网络Socket端口,Spark Streaming连接上该端口并在运行的时候一直监听该端口

                     *                  的数据(当然该端口服务首先必须存在),并且在后续会根据业务需要不断的有数据产生(当然对于Spark Streaming

                     *                  应用程序的运行而言,有无数据其处理流程都是一样的);

                     * 3,如果经常在每间隔5秒钟没有数据的话不断的启动空的Job其实是会造成调度资源的浪费,因为并没有数据需要发生计算,所以

                     *                 实例的企业级生成环境的代码在具体提交Job前会判断是否有数据,如果没有的话就不再提交Job;

                     */                 

                    JavaReceiverInputDStream lines = FlumeUtils.createPollingStream(jsc, "Master", 9999);           

                    /*

                     * 第四步:接下来就像对于RDD编程一样基于DStream进行编程!!!原因是DStream是RDD产生的模板(或者说类),在Spark Streaming具体

                     * 发生计算前,其实质是把每个Batch的DStream的操作翻译成为对RDD的操作!!!

                     *对初始的DStream进行Transformation级别的处理,例如map、filter等高阶函数等的编程,来进行具体的数据计算

               *    第4.1步:讲每一行的字符串拆分成单个的单词

               */

                    JavaDStream words = lines.flatMap(new FlatMapFunction() { //如果是Scala,由于SAM转换,所以可以写成val words = lines.flatMap { line => line.split(" ")}

                               @Override

                               public Iterable call(SparkFlumeEvent event) throws Exception {

                                         String line = new String(event.event().getBody().array());

                                         return Arrays.asList(line.split(" "));

                               }

                    });

                     /*

                 * 第四步:对初始的DStream进行Transformation级别的处理,例如map、filter等高阶函数等的编程,来进行具体的数据计算

                 * 第4.2步:在单词拆分的基础上对每个单词实例计数为1,也就是word => (word, 1)

                 */

                    JavaPairDStream pairs = words.mapToPair(new PairFunction() {

                               @Override

                               public Tuple2 call(String word) throws Exception {

                                         return new Tuple2(word, 1);

                               }

                    });           

                     /*

                 * 第四步:对初始的DStream进行Transformation级别的处理,例如map、filter等高阶函数等的编程,来进行具体的数据计算

                 * 第4.3步:在每个单词实例计数为1基础之上统计每个单词在文件中出现的总次数

                 */

                    JavaPairDStream wordsCount = pairs.reduceByKey(new Function2() { //对相同的Key,进行Value的累计(包括Local和Reducer级别同时Reduce)     

                               @Override

                               public Integer call(Integer v1, Integer v2) throws Exception {

                                         return v1 + v2;

                               }

                    });

                    /*

                     *        此处的print并不会直接出发Job的执行,因为现在的一切都是在Spark Streaming框架的控制之下的,对于Spark Streaming

                     *        而言具体是否触发真正的Job运行是基于设置的Duration时间间隔的

                     *        诸位一定要注意的是Spark Streaming应用程序要想执行具体的Job,对Dtream就必须有output Stream操作,

                     *        output Stream有很多类型的函数触发,类print、saveAsTextFile、saveAsHadoopFiles等,最为重要的一个

                     *        方法是foraeachRDD,因为Spark Streaming处理的结果一般都会放在Redis、DB、DashBoard等上面,foreachRDD

                     *        主要就是用用来完成这些功能的,而且可以随意的自定义具体数据到底放在哪里!!!

                     *

                     */

                    wordsCount.print();

                    /*

                     * Spark Streaming执行引擎也就是Driver开始运行,Driver启动的时候是位于一条新的线程中的,当然其内部有消息循环体,用于

                     * 接受应用程序本身或者Executor中的消息;

                     */

                    jsc.start();                

                    jsc.awaitTermination();

                    jsc.close();

          }

启动HDFS集群:

启动运行Flume:

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第5张图片

启动eclipse下的应用程序:

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第6张图片

copy测试文件hellospark.txt到Flume flume-conf.properties配置文件中指定的/usr/local/flume/tmp/TestDir目录下:

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第7张图片

Spark Streaming从Flume Poll数据案例实战和内幕源码解密_第8张图片

隔30秒后可以在eclipse程序控制台中看到上传的文件单词统计结果。

第二部分:源码分析

1、创建createPollingStream (FlumeUtils.scala )

注意:默认的存储方式是MEMORY_AND_DISK_SER_2

/**

 * Creates an input stream that is to be used with the Spark Sink deployed on a Flume agent.

 * This stream will poll the sink for data and will pull events as they are available.

 * This stream will use a batch size of 1000 events and run 5 threads to pull data.

 * @param hostname Address of the host on which the Spark Sink is running

 * @param port Port of the host at which the Spark Sink is listening

 * @param storageLevel Storage level to use for storing the received objects

*/

def createPollingStream(

    ssc: StreamingContext,

    hostname: String,

    port: Int,

    storageLevel: StorageLevel = StorageLevel.MEMORY_AND_DISK_SER_2

  ): ReceiverInputDStream[SparkFlumeEvent] = {

  createPollingStream(ssc, Seq(new InetSocketAddress(hostname, port)), storageLevel)

}

2、参数配置:默认的全局参数,private 级别配置无法修改

private val DEFAULT_POLLING_PARALLELISM = 5

private val DEFAULT_POLLING_BATCH_SIZE = 1000

/**

 * Creates an input stream that is to be used with the Spark Sink deployed on a Flume agent.

 * This stream will poll the sink for data and will pull events as they are available.

 * This stream will use a batch size of 1000 events and run 5 threads to pull data.

 * @param addresses List of InetSocketAddresses representing the hosts to connect to.

 * @param storageLevel Storage level to use for storing the received objects

 */

def createPollingStream(

    ssc: StreamingContext,

    addresses: Seq[InetSocketAddress],

    storageLevel: StorageLevel

  ): ReceiverInputDStream[SparkFlumeEvent] = {

  createPollingStream(ssc, addresses, storageLevel,

    DEFAULT_POLLING_BATCH_SIZE, DEFAULT_POLLING_PARALLELISM)

}

3、创建FlumePollingInputDstream对象

/**

 * Creates an input stream that is to be used with the Spark Sink deployed on a Flume agent.

 * This stream will poll the sink for data and will pull events as they are available.

 * @param addresses List of InetSocketAddresses representing the hosts to connect to.

 * @param maxBatchSize Maximum number of events to be pulled from the Spark sink in a

 *                     single RPC call

 * @param parallelism Number of concurrent requests this stream should send to the sink. Note

 *                    that having a higher number of requests concurrently being pulled will

 *                    result in this stream using more threads

 * @param storageLevel Storage level to use for storing the received objects

 */

def createPollingStream(

    ssc: StreamingContext,

    addresses: Seq[InetSocketAddress],

    storageLevel: StorageLevel,

    maxBatchSize: Int,

    parallelism: Int

  ): ReceiverInputDStream[SparkFlumeEvent] = {

  new FlumePollingInputDStream[SparkFlumeEvent](ssc, addresses, maxBatchSize,

    parallelism, storageLevel)

}

4、继承自ReceiverInputDstream并覆写getReciver方法,调用FlumePollingReciver接口

private[streaming] class FlumePollingInputDStream[T: ClassTag](

    _ssc: StreamingContext,

    val addresses: Seq[InetSocketAddress],

    val maxBatchSize: Int,

    val parallelism: Int,

    storageLevel: StorageLevel

  ) extends ReceiverInputDStream[SparkFlumeEvent](_ssc) {

   override def getReceiver(): Receiver[SparkFlumeEvent] = {

    new FlumePollingReceiver(addresses, maxBatchSize, parallelism, storageLevel)

  }

}

5、ReceiverInputDstream 构建了一个线程池,设置为后台线程;并使用lazy和工厂方法创建线程和NioClientSocket(NioClientSocket底层使用NettyServer的方式)

lazy val channelFactoryExecutor =

  Executors.newCachedThreadPool(new ThreadFactoryBuilder().setDaemon(true).

    setNameFormat("Flume Receiver Channel Thread - %d").build())

lazy val channelFactory =

  new NioClientSocketChannelFactory(channelFactoryExecutor, channelFactoryExecutor)

6、receiverExecutor 内部也是线程池;connections是指链接分布式Flume集群的FlumeConnection实体句柄的个数,线程拿到实体句柄访问数据。

lazy val receiverExecutor = Executors.newFixedThreadPool(parallelism,

  new ThreadFactoryBuilder().setDaemon(true).setNameFormat("Flume Receiver Thread - %d").build())

private lazy val connections = new LinkedBlockingQueue[FlumeConnection]()

7、启动时创建NettyTransceiver,根据并行度(默认5个)循环提交FlumeBatchFetcher

override def onStart(): Unit = {

  // Create the connections to each Flume agent.

  addresses.foreach(host => {

    val transceiver = new NettyTransceiver(host, channelFactory)

    val client = SpecificRequestor.getClient(classOf[SparkFlumeProtocol.Callback], transceiver)

    connections.add(new FlumeConnection(transceiver, client))

  })

  for (i <- 0 until parallelism) {

    logInfo("Starting Flume Polling Receiver worker threads..")

    // Threads that pull data from Flume.

    receiverExecutor.submit(new FlumeBatchFetcher(this))

  }

}

8、FlumeBatchFetcher run方法中从Receiver中获取connection链接句柄ack跟消息确认有关

def run(): Unit = {

  while (!receiver.isStopped()) {

    val connection = receiver.getConnections.poll()

    val client = connection.client

    var batchReceived = false

    var seq: CharSequence = null

    try {

      getBatch(client) match {

        case Some(eventBatch) =>

          batchReceived = true

          seq = eventBatch.getSequenceNumber

          val events = toSparkFlumeEvents(eventBatch.getEvents)

          if (store(events)) {

            sendAck(client, seq)

          } else {

            sendNack(batchReceived, client, seq)

          }

        case None =>

      }

    } catch {

9、获取一批一批数据方法

/**

 * Gets a batch of events from the specified client. This method does not handle any exceptions

 * which will be propogated to the caller.

 * @param client Client to get events from

 * @return [[Some]] which contains the event batch if Flume sent any events back, else [[None]]

 */

private def getBatch(client: SparkFlumeProtocol.Callback): Option[EventBatch] = {

  val eventBatch = client.getEventBatch(receiver.getMaxBatchSize)

  if (!SparkSinkUtils.isErrorBatch(eventBatch)) {

    // No error, proceed with processing data

    logDebug(s"Received batch of ${eventBatch.getEvents.size} events with sequence " +

      s"number: ${eventBatch.getSequenceNumber}")

    Some(eventBatch)

  } else {

    logWarning("Did not receive events from Flume agent due to error on the Flume agent: " +

      eventBatch.getErrorMsg)

    None

  }

}

总结:

88课

备注:

资料来源于:DT_大数据梦工厂(IMF传奇行动绝密课程)

更多私密内容,请关注微信公众号:DT_Spark

转载于:https://www.cnblogs.com/sparkbigdata/p/5448673.html

你可能感兴趣的:(Spark Streaming从Flume Poll数据案例实战和内幕源码解密)