为什么阿里不允许用Executors创建线程池

1 文章概述

《阿里巴巴JAVA开发手册》有这样一条强制规定：线程池不允许使用Executors去创建，而应该通过ThreadPoolExecutor方式，这样处理方式更加明确线程池运行规则，规避资源耗尽风险。本文我们从资源和排查问题两个角度进行分析，同时参考DUBBO线程池声明方式创建一个符合规范的线程池。

2 资源角度

《阿里巴巴JAVA开发手册》从资源角度对这个问题进行了分析

FixedThreadPool SingleThreadPool
允许请求队列长度为Integer.MAX_VALUE可能会堆积大量请求从而导致OOM


CachedThreadPool ScheduledThreadPool
允许创建线程数量为Integer.MAX_VALUE可能会创建大量线程从而导致OOM

以下两个线程池使用链表实现的阻塞队列，不设大小理论上队列容量无上限，所以可能会堆积大量请求从而导致OOM

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


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

以下两个线程池maxSize使用Integer最大值，所以可能会创建大量线程从而导致OOM

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


public static ScheduledExecutorService newSingleThreadScheduledExecutor() {
    return new DelegatedScheduledExecutorService
           (new ScheduledThreadPoolExecutor(1));
}


public ScheduledThreadPoolExecutor(int corePoolSize) {
    super(corePoolSize, Integer.MAX_VALUE, 0, NANOSECONDS,
          new DelayedWorkQueue());
}

3 排查问题角度

如果使用Executors创建线程池，大家应该最常使用如下语句

public void testThread() throws Exception {
    ExecutorService fixedExecutor = Executors.newFixedThreadPool(10);
    for (int i = 0; i < 1000; i++) {
        fixedExecutor.execute(new Runnable() {
            @Override
            public void run() {
                System.out.println("公众号互联网公园");
            }
        });
    }
}

上述语句在功能层面是没有问题的，但是在生产环境中有可能遇到CPU飙高，线程数持续增加，内存溢出等问题，我们时常需要通过线程快照进行观察。我们通过jstack命令观察上述代码线程快照

"pool-1-thread-2" #525 prio=5 os_prio=0 tid=0x00006f6561039100 nid=0xdaa waiting on condition [0x00006f64e646d000]
java.lang.Thread.State: WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
parking to wait for <0x00000006e6f3e230> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java:165)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1066)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1126)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:616)
at java.lang.Thread.run(Thread.java:645)

我们发现从线程快照看不出任何业务信息，只有类似pool-1-thread-2这种编号信息，不利于排查问题，我们需要给线程命名。

4 为线程进行命名

在并发编程中我们一定要为线程命名，这样有助于排查问题，关于如何命名我们可以参考DUBBO源码，分析FixedThreadPool线程池会发现其使用命名工厂为生产者和消费者线程进行命名

public class FixedThreadPool implements ThreadPool {


    @Override
    public Executor getExecutor(URL url) {


        // 线程名称
        String name = url.getParameter(Constants.THREAD_NAME_KEY, Constants.DEFAULT_THREAD_NAME);


        // 线程个数默认200
        int threads = url.getParameter(Constants.THREADS_KEY, Constants.DEFAULT_THREADS);


        // 队列容量默认0
        int queues = url.getParameter(Constants.QUEUES_KEY, Constants.DEFAULT_QUEUES);


        // 队列容量等于0使用阻塞队列SynchronousQueue
        // 队列容量小于0使用无界阻塞队列LinkedBlockingQueue
        // 队列容量大于0使用有界阻塞队列LinkedBlockingQueue
        // NamedInternalThreadFactory为线程命名
        return new ThreadPoolExecutor(threads, threads, 0, TimeUnit.MILLISECONDS,
                                      queues == 0 ? new SynchronousQueue<Runnable>()
                                      : (queues < 0 ? new LinkedBlockingQueue<Runnable>()
                                         : new LinkedBlockingQueue<Runnable>(queues)),
                                      new NamedInternalThreadFactory(name, true), new AbortPolicyWithReport(name, url));
    }
}

生产者默认线程名DubboServerHandler

public abstract class AbstractServer extends AbstractEndpoint implements Server {
    protected static final String SERVER_THREAD_POOL_NAME = "DubboServerHandler";
}

生产者线程快照信息如下

"DubboServerHandler-1.1.1.1:20881-thread-20" #511 daemon prio = 5 os_prio = 0 tid = 0x00001f153121f200 nid = 0xd1a waiting on condition [0x00001f14edcdf000]
java.lang.Thread.State: WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x00000001e1f3abc0> (a java.util.concurrent.SynchronousQueue$TransferStack)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java : 115)
at java.util.concurrent.SynchronousQueue$TransferStack.awaitFulfill(SynchronousQueue.java : 452)
at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java : 312)
at java.util.concurrent.SynchronousQueue.take(SynchronousQueue.java : 924)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java : 1011)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java : 1121)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java : 111)

消费者默认线程名DubboClientHandler

public abstract class AbstractClient extends AbstractEndpoint implements Client {
    protected static final String CLIENT_THREAD_POOL_NAME = "DubboClientHandler";
}

消费者线程快照信息如下

"DubboClientHandler-1.1.1.1:20881-thread-10" #688 daemon prio=1 os_prio=0 tid=0x00001f6114004800 nid=0x14d8 waiting on condition [0x00001f63e131a000]
java.lang.Thread.State: TIMED_WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x00000006e21df0d0> (a java.util.concurrent.SynchronousQueue$TransferStack)
at java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:111)
at java.util.concurrent.SynchronousQueue$TransferStack.awaitFulfill(SynchronousQueue.java:460)
at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java:361)
at java.util.concurrent.SynchronousQueue.poll(SynchronousQueue.java:141)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1066)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1111)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:611)
at java.lang.Thread.run(Thread.java:141)

5 创建一个规范线程池

我们参考DUBBO线程池定义命名工厂

public class NamedInternalThreadFactory extends NamedThreadFactory {
    public NamedInternalThreadFactory() {
        super();
    }


    public NamedInternalThreadFactory(String prefix) {
        super(prefix, false);
    }


    public NamedInternalThreadFactory(String prefix, boolean daemon) {
        super(prefix, daemon);
    }


    @Override
    public Thread newThread(Runnable runnable) {
        String name = mPrefix + mThreadNum.getAndIncrement();
        InternalThread ret = new InternalThread(mGroup, runnable, name, 0);
        ret.setDaemon(mDaemon);
        return ret;
    }
}


public class NamedThreadFactory implements ThreadFactory {
    protected static final AtomicInteger POOL_SEQ = new AtomicInteger(1);
    protected final AtomicInteger mThreadNum = new AtomicInteger(1);
    protected final String mPrefix;
    protected final boolean mDaemon;
    protected final ThreadGroup mGroup;


    public NamedThreadFactory() {
        this("pool-" + POOL_SEQ.getAndIncrement(), false);
    }


    public NamedThreadFactory(String prefix) {
        this(prefix, false);
    }


    public NamedThreadFactory(String prefix, boolean daemon) {
        mPrefix = prefix + "-thread-";
        mDaemon = daemon;
        SecurityManager s = System.getSecurityManager();
        mGroup = (s == null) ? Thread.currentThread().getThreadGroup() : s.getThreadGroup();
    }


    @Override
    public Thread newThread(Runnable runnable) {
        String name = mPrefix + mThreadNum.getAndIncrement();
        Thread ret = new Thread(mGroup, runnable, name, 0);
        ret.setDaemon(mDaemon);
        return ret;
    }


    public ThreadGroup getThreadGroup() {
        return mGroup;
    }
}

再定义一个线程池，在线程池执行方法开放一个业务名称参数供调用方设置

public class ThreadPoolStarter {
      public static ThreadPoolExecutor getExecutor(String threadName) {
        if (executor == null) {
          synchronized (ThreadPoolStarter.class) {
            if (executor == null) {
              int coreSize = Runtime.getRuntime().availableProcessors();
              BlockingQueue<Runnable> queueToUse = new LinkedBlockingQueue<Runnable>(QUEUE_SIZE);
              executor = new ThreadPoolExecutor(coreSize, POOL_CORE_SIZE, MAX_SIZE, TimeUnit.SECONDS, queueToUse, new NamedInternalThreadFactory(threadName, true), new AbortPolicyDoReport(threadName));
            }
      }
    }
    return executor;
  }
}


public class ThreadExecutor {
  public static void execute(String bizName, Runnable job) {
    ThreadPoolStarter.getExecutor(bizName).execute(job);
  }


  public static Future<?> sumbit(String bizName, Runnable job) {
    return ThreadPoolStarter.getExecutor(bizName).submit(job);
  }
}

编写一个实例进行测试

public void testThread() throws Exception {
    for (int i = 0; i < 10000; i++) {
      ThreadExecutor.execute("BizName", new Runnable() {
        @Override
        public void run() {
          System.out.println("公众号互联网公园");
        }
      });
      Thread.sleep(1000L);
    }
  }
}

再观察线程快照可以清晰查看业务名

"BizName-thread-8" #262 daemon prio=5 os_prio=0 tid=0x0000000023b5c000 nid=0x31d4 waiting on condition [0x000000003c0be000]
java.lang.Thread.State: WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x00000006c35781f0> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)

6 文章总结

本文首先介绍了《阿里巴巴JAVA开发手册》不允许使用Executors创建线程池这个规定，然后从资源和排查问题两个角度分析了为什么这么规定，最后我们参考DUBBO线程池声明方式创建了一个规范线程池，这样使用线程池有助于快速定位和排查问题。

特别推荐一个分享架构+算法的优质内容，还没关注的小伙伴，可以长按关注一下：

长按订阅更多精彩▼

如有收获，点个在看，诚挚感谢