Spark源码学习--内置RPC框架(2)

RPC配置类 TransportConf

TransportConf给Spark的RPC框架提供配置信息，它有两个成员属性——配置提供者conf和配置的模块名称module。这两个属性的定义如下：

//配置提供者
private final ConfigProvider conf;
//模块名称
private final String module;

ConfigProvider是一个抽象类，代码如下：

/**
 * Provides a mechanism for constructing a {@link TransportConf} using some sort of configuration.
 */
public abstract class ConfigProvider {
  /** Obtains the value of the given config, throws NoSuchElementException if it doesn't exist. */
  public abstract String get(String name);

  public String get(String name, String defaultValue) {
    try {
      return get(name);
    } catch (NoSuchElementException e) {
      return defaultValue;
    }
  }

  public int getInt(String name, int defaultValue) {
    return Integer.parseInt(get(name, Integer.toString(defaultValue)));
  }

  public long getLong(String name, long defaultValue) {
    return Long.parseLong(get(name, Long.toString(defaultValue)));
  }

  public double getDouble(String name, double defaultValue) {
    return Double.parseDouble(get(name, Double.toString(defaultValue)));
  }

  public boolean getBoolean(String name, boolean defaultValue) {
    return Boolean.parseBoolean(get(name, Boolean.toString(defaultValue)));
  }
}

ConfigProvider中包括get、getInt、getLong、getDouble、getBoolean等方法，这些方法都是基于抽象方法get获取值，经过一次类型转换而实现。这个抽象的get方法将需要子类去实现。

实际代码中，get的实现实际是代理了SparkConf的get方法

/** Get a parameter; throws a NoSuchElementException if it's not set */
  def get(key: String): String = {
    getOption(key).getOrElse(throw new NoSuchElementException(key))
  }

由scala单例类SparkTransportConf创建TransportConf对象。

package org.apache.spark.network.netty

import org.apache.spark.SparkConf
import org.apache.spark.network.util.{ConfigProvider, TransportConf}

/**
 * Provides a utility for transforming from a SparkConf inside a Spark JVM (e.g., Executor,
 * Driver, or a standalone shuffle service) into a TransportConf with details on our environment
 * like the number of cores that are allocated to this JVM.
 */
object SparkTransportConf {
  /**
   * Specifies an upper bound on the number of Netty threads that Spark requires by default.
   * In practice, only 2-4 cores should be required to transfer roughly 10 Gb/s, and each core
   * that we use will have an initial overhead of roughly 32 MB of off-heap memory, which comes
   * at a premium.
   *
   * Thus, this value should still retain maximum throughput and reduce wasted off-heap memory
   * allocation. It can be overridden by setting the number of serverThreads and clientThreads
   * manually in Spark's configuration.
   */
  private val MAX_DEFAULT_NETTY_THREADS = 8

  /**
   * Utility for creating a [[TransportConf]] from a [[SparkConf]].
   * @param _conf the [[SparkConf]]
   * @param module the module name
   * @param numUsableCores if nonzero, this will restrict the server and client threads to only
   *                       use the given number of cores, rather than all of the machine's cores.
   *                       This restriction will only occur if these properties are not already set.
   */
  def fromSparkConf(_conf: SparkConf, module: String, numUsableCores: Int = 0): TransportConf = {
    val conf = _conf.clone

    // Specify thread configuration based on our JVM's allocation of cores (rather than necessarily
    // assuming we have all the machine's cores).
    // NB: Only set if serverThreads/clientThreads not already set.
    val numThreads = defaultNumThreads(numUsableCores)
    conf.setIfMissing(s"spark.$module.io.serverThreads", numThreads.toString)
    conf.setIfMissing(s"spark.$module.io.clientThreads", numThreads.toString)

    new TransportConf(module, new ConfigProvider {
      override def get(name: String): String = conf.get(name)
    })
  }

  /**
   * Returns the default number of threads for both the Netty client and server thread pools.
   * If numUsableCores is 0, we will use Runtime get an approximate number of available cores.
   */
  private def defaultNumThreads(numUsableCores: Int): Int = {
    val availableCores =
      if (numUsableCores > 0) numUsableCores else Runtime.getRuntime.availableProcessors()
    math.min(availableCores, MAX_DEFAULT_NETTY_THREADS)
  }
}

从代码看到，可以使用SparkTransportConf的fromSparkConf方法来构造TransportConf。传递的三个参数分别为SparkConf、模块名module及可用的内核数num-UsableCores。如果numUsableCores小于等于0，那么线程数是系统可用处理器的数量，不过系统的内核数不可能全部用于网络传输，所以这里将分配给网络传输的内核数量最多限制在8个。

最终确定的线程数将用于设置客户端传输线程spark.module.io.clientThreads属性和服务端传输线程数spark.module.io.serverThreads属性。

                           博客基于《Spark内核设计的艺术：架构设计与实现》一书