Java 源码 - java.util.concurrent.Executors

线程池的定义

管理一组工作线程。通过线程池复用线程有以下几点优点：

• 减少资源创建 => 减少内存开销，创建线程占用内存
• 降低系统开销 => 创建线程需要时间，会延迟处理的请求
• 提高稳定稳定性 => 避免无限创建线程引起的OutOfMemoryError【简称OOM】

Executors的构造方法

    /** Cannot instantiate. */
    private Executors() {}

Executors创建线程池的方式

根据返回的对象类型创建线程池可以分为三类：

• 创建返回ThreadPoolExecutor对象
• 创建返回ScheduleThreadPoolExecutor对象
• 创建返回ForkJoinPool对象

本文主要讨论创建返回ThreadPoolExecutor对象

ThreadPoolExecutor对象

在介绍Executors创建线程池方法前先介绍一下ThreadPoolExecutor，因为这些创建线程池的静态方法都是返回ThreadPoolExecutor对象，和我们手动创建ThreadPoolExecutor对象的区别就是我们不需要自己传构造函数的参数。

ThreadPoolExecutor的构造函数共有四个，但最终调用的都是同一个：

public ThreadPoolExecutor(int corePoolSize,
                          int maximumPoolSize,
                          long keepAliveTime,
                          TimeUnit unit,
                          BlockingQueue<Runnable> workQueue,
                          ThreadFactory threadFactory,
                          RejectedExecutionHandler handler)

• corePoolSize => 线程池核心线程数量
• maximumPoolSize => 线程池最大数量
• keepAliveTime => 空闲线程存活时间
• unit => 时间单位
• workQueue => 线程池所使用的缓冲队列
• threadFactory => 线程池创建线程使用的工厂
• handler => 线程池对拒绝任务的处理策略

RejectedExecutionHandler ： 饱和策略。这是当任务队列和线程数都满了的时候所采取的的对应策略默认是AbordPolicy

AbordPolicy:   表示无法处理新的任务，并抛出RejectedExecutionException异常。

CallerRunsPolicy ： 用调用者所在的线程来处理任务。

DisCardPolicy ： 不能执行任务并将任务删除。

DisCardOldesPolicy ： 丢弃队列最近的任务，并执行当前的任务， 会一直执行下去。

线程池执行任务逻辑和线程池参数的关系

执行逻辑说明：

• 判断核心线程数是否已满，核心线程数大小和corePoolSize参数有关，未满则创建线程执行任务
• 若核心线程池已满，判断队列是否满，队列是否满和workQueue参数有关，若未满则加入队列中
• 若队列已满，判断线程池是否已满，线程池是否已满和maximumPoolSize参数有关，若未满创建线程执行任务
• 若线程池已满，则采用拒绝策略处理无法执执行的任务，拒绝策略和handler参数有关

Executors创建返回ThreadPoolExecutor对象

Executors创建返回ThreadPoolExecutor对象的方法共有三种：

• Executors#newCachedThreadPool => 创建可缓存的线程池
• Executors#newSingleThreadExecutor => 创建单线程的线程池
• Executors#newFixedThreadPool => 创建固定长度的线程池

Executors#newCachedThreadPool方法

public static ExecutorService newCachedThreadPool() {
    return new ThreadPoolExecutor(0, Integer.MAX_VALUE,
                                  60L, TimeUnit.SECONDS,
                                  new SynchronousQueue<Runnable>());
}

• corePoolSize => 0，核心线程池的数量为0
• maximumPoolSize => Integer.MAX_VALUE，可以认为最大线程数是无限的
• keepAliveTime => 60L
• unit => 秒
• workQueue => SynchronousQueue

当一个任务提交时，corePoolSize为0不创建核心线程，SynchronousQueue是一个不存储元素的队列，可以理解为队里永远是满的，因此最终会创建非核心线程来执行任务。

对于非核心线程空闲60s时将被回收。因为Integer.MAX_VALUE非常大，可以认为是可以无限创建线程的，在资源有限的情况下容易引起OOM异常

Executors#newSingleThreadExecutor方法

public static ExecutorService newSingleThreadExecutor() {
    return new FinalizableDelegatedExecutorService
        (new ThreadPoolExecutor(1, 1,
                                0L, TimeUnit.MILLISECONDS,
                                new LinkedBlockingQueue<Runnable>()));
}

• corePoolSize => 1，核心线程池的数量为1
• maximumPoolSize => 1，只可以创建一个非核心线程
• keepAliveTime => 0L
• unit => 秒
• workQueue => LinkedBlockingQueue

当一个任务提交时，首先会创建一个核心线程来执行任务，如果超过核心线程的数量，将会放入队列中，因为LinkedBlockingQueue是长度为Integer.MAX_VALUE的队列，可以认为是无界队列，因此往队列中可以插入无限多的任务，在资源有限的时候容易引起OOM异常，同时因为无界队列，maximumPoolSize和keepAliveTime参数将无效，压根就不会创建非核心线程

Executors#newFixedThreadPool方法

public static ExecutorService newFixedThreadPool(int nThreads) {
    return new ThreadPoolExecutor(nThreads, nThreads,
                                  0L, TimeUnit.MILLISECONDS,
                                  new LinkedBlockingQueue<Runnable>());
}

• corePoolSize => 1，核心线程池的数量为1
• maximumPoolSize => 1，只可以创建一个非核心线程
• keepAliveTime => 0L
• unit => 秒
• workQueue => LinkedBlockingQueue
• 它和SingleThreadExecutor类似，唯一的区别就是核心线程数不同，并且由于使用的是LinkedBlockingQueue，在资源有限的时候容易引起OOM异常

Executors创建返回ForkJoinPool对象

    /**
     * Creates a thread pool that maintains enough threads to support
     * the given parallelism level, and may use multiple queues to
     * reduce contention. The parallelism level corresponds to the
     * maximum number of threads actively engaged in, or available to
     * engage in, task processing. The actual number of threads may
     * grow and shrink dynamically. A work-stealing pool makes no
     * guarantees about the order in which submitted tasks are
     * executed.
     *
     * @param parallelism the targeted parallelism level
     * @return the newly created thread pool
     * @throws IllegalArgumentException if {@code parallelism <= 0}
     * @since 1.8
     */
    public static ExecutorService newWorkStealingPool(int parallelism) {
        return new ForkJoinPool
            (parallelism,
             ForkJoinPool.defaultForkJoinWorkerThreadFactory,
             null, true);
    }

    /**
     * Creates a work-stealing thread pool using all
     * {@link Runtime#availableProcessors available processors}
     * as its target parallelism level.
     * @return the newly created thread pool
     * @see #newWorkStealingPool(int)
     * @since 1.8
     */
    public static ExecutorService newWorkStealingPool() {
        return new ForkJoinPool
            (Runtime.getRuntime().availableProcessors(),
             ForkJoinPool.defaultForkJoinWorkerThreadFactory,
             null, true);
    }

Executors创建返回ScheduledExecutorService对象

    /**
     * Creates a single-threaded executor that can schedule commands
     * to run after a given delay, or to execute periodically.
     * (Note however that if this single
     * thread terminates due to a failure during execution prior to
     * shutdown, a new one will take its place if needed to execute
     * subsequent tasks.)  Tasks are guaranteed to execute
     * sequentially, and no more than one task will be active at any
     * given time. Unlike the otherwise equivalent
     * {@code newScheduledThreadPool(1)} the returned executor is
     * guaranteed not to be reconfigurable to use additional threads.
     * @return the newly created scheduled executor
     */
    public static ScheduledExecutorService newSingleThreadScheduledExecutor() {
        return new DelegatedScheduledExecutorService
            (new ScheduledThreadPoolExecutor(1));
    }

    /**
     * Creates a single-threaded executor that can schedule commands
     * to run after a given delay, or to execute periodically.  (Note
     * however that if this single thread terminates due to a failure
     * during execution prior to shutdown, a new one will take its
     * place if needed to execute subsequent tasks.)  Tasks are
     * guaranteed to execute sequentially, and no more than one task
     * will be active at any given time. Unlike the otherwise
     * equivalent {@code newScheduledThreadPool(1, threadFactory)}
     * the returned executor is guaranteed not to be reconfigurable to
     * use additional threads.
     * @param threadFactory the factory to use when creating new
     * threads
     * @return a newly created scheduled executor
     * @throws NullPointerException if threadFactory is null
     */
    public static ScheduledExecutorService newSingleThreadScheduledExecutor(ThreadFactory threadFactory) {
        return new DelegatedScheduledExecutorService
            (new ScheduledThreadPoolExecutor(1, threadFactory));
    }

    /**
     * Creates a thread pool that can schedule commands to run after a
     * given delay, or to execute periodically.
     * @param corePoolSize the number of threads to keep in the pool,
     * even if they are idle
     * @return a newly created scheduled thread pool
     * @throws IllegalArgumentException if {@code corePoolSize < 0}
     */
    public static ScheduledExecutorService newScheduledThreadPool(int corePoolSize) {
        return new ScheduledThreadPoolExecutor(corePoolSize);
    }

    /**
     * Creates a thread pool that can schedule commands to run after a
     * given delay, or to execute periodically.
     * @param corePoolSize the number of threads to keep in the pool,
     * even if they are idle
     * @param threadFactory the factory to use when the executor
     * creates a new thread
     * @return a newly created scheduled thread pool
     * @throws IllegalArgumentException if {@code corePoolSize < 0}
     * @throws NullPointerException if threadFactory is null
     */
    public static ScheduledExecutorService newScheduledThreadPool(
            int corePoolSize, ThreadFactory threadFactory) {
        return new ScheduledThreadPoolExecutor(corePoolSize, threadFactory);
    }

Executors创建返回Callable对象

    public static <T> Callable<T> callable(Runnable task, T result) {
        if (task == null)
            throw new NullPointerException();
        return new RunnableAdapter<T>(task, result);
    }    

    public static Callable<Object> callable(Runnable task) {
        if (task == null)
            throw new NullPointerException();
        return new RunnableAdapter<Object>(task, null);
    }

Executors静态类RunnableAdapter

    /**
     * A callable that runs given task and returns given result
     */
    static final class RunnableAdapter<T> implements Callable<T> {
        final Runnable task;
        final T result;
        RunnableAdapter(Runnable task, T result) {
            this.task = task;
            this.result = result;
        }
        public T call() {
            task.run();
            return result;
        }
    }

Executors静态类DefaultThreadFactory

    /**
     * The default thread factory
     */
    static class DefaultThreadFactory implements ThreadFactory {
        private static final AtomicInteger poolNumber = new AtomicInteger(1);
        private final ThreadGroup group;
        private final AtomicInteger threadNumber = new AtomicInteger(1);
        private final String namePrefix;

        DefaultThreadFactory() {
            SecurityManager s = System.getSecurityManager();
            group = (s != null) ? s.getThreadGroup() :
                                  Thread.currentThread().getThreadGroup();
            namePrefix = "pool-" +
                          poolNumber.getAndIncrement() +
                         "-thread-";
        }

        public Thread newThread(Runnable r) {
            Thread t = new Thread(group, r,
                                  namePrefix + threadNumber.getAndIncrement(),
                                  0);
            if (t.isDaemon())
                t.setDaemon(false);
            if (t.getPriority() != Thread.NORM_PRIORITY)
                t.setPriority(Thread.NORM_PRIORITY);
            return t;
        }
    }

总结：

• FixedThreadPool和SingleThreadExecutor => 允许的请求队列长度为Integer.MAX_VALUE，可能会堆积大量的请求，从而引起OOM异常
• CachedThreadPool => 允许创建的线程数为Integer.MAX_VALUE，可能会创建大量的线程，从而引起OOM异常

这就是为什么禁止使用Executors去创建线程池，而是推荐自己去创建ThreadPoolExecutor的原因

OOM异常测试

理论上会出现OOM异常，必须测试一波验证之前的说法：

测试类：TaskTest.java

public class TaskTest {
    public static void main(String[] args) {
        ExecutorService es = Executors.newCachedThreadPool();
        int i = 0;
        while (true) {
            es.submit(new Task(i++));
        }
    }
}

在启动测试类之前先将JVM内存调整小一点。

JVM参数说明：

• -Xms10M => Java Heap内存初始化值
• -Xmx10M => Java Heap内存最大值

运行结果：

Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "main"
Disconnected from the target VM, address: '127.0.0.1:60416', transport: 'socket'

另外两个线程池就不做测试了，测试方法一致，只是创建的线程池不一样

如何定义线程池参数

CPU密集型 => 线程池的大小推荐为CPU数量 + 1，CPU数量可以根据Runtime.availableProcessors方法获取

IO密集型 => CPU数量

(1 + 线程等待时间/线程CPU时间)

混合型 => 将任务分为CPU密集型和IO密集型，然后分别使用不同的线程池去处理，从而使每个线程池可以根据各自的工作负载来调整

阻塞队列 => 推荐使用有界队列，有界队列有助于避免资源耗尽的情况发生

拒绝策略 => 默认采用的是AbortPolicy拒绝策略，直接在程序中抛出RejectedExecutionException异常【因为是运行时异常，不强制catch】，这种处理方式不够优雅。处理拒绝策略有以下几种比较推荐：

• 在程序中捕获RejectedExecutionException异常，在捕获异常中对任务进行处理。针对默认拒绝策略
• 使用CallerRunsPolicy拒绝策略，该策略会将任务交给调用execute的线程执行【一般为主线程】，此时主线程将在一段时间内不能提交任何任务，从而使工作线程处理正在执行的任务。此时提交的线程将被保存在TCP队列中，TCP队列满将会影响客户端，这是一种平缓的性能降低
• 自定义拒绝策略，只需要实现RejectedExecutionHandler接口即可
• 如果任务不是特别重要，使用DiscardPolicy和DiscardOldestPolicy拒绝策略将任务丢弃也是可以的

如果使用Executors的静态方法创建ThreadPoolExecutor对象，可以通过使用Semaphore对任务的执行进行限流也可以避免出现OOM异常