为何阿里巴巴不容许用Executors建立线程池

时间  2021-08-14
标签 java 编程 微信 markdown 架构 并发 ide 源码分析 学习 测试 栏目 阿里巴巴 繁體版
原文   原文链接

欢迎你们关注公众号「JAVA前线」查看更多精彩分享文章，主要包括源码分析、实际应用、架构思惟、职场分享、产品思考等等，同时欢迎你们加我微信「java_front」一块儿交流学习java


1 文章概述

阿里开发手册有这样一条强制规定：线程池不容许使用Executors去建立，而应该经过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。本文咱们从资源和排查问题两个角度进行分析，同时参考DUBBO线程池声明方式建立一个符合规范的线程池。编程


2 资源角度

阿里巴巴JAVA开发手册主要是从资源角度对这个问题进行了分析：微信

FixedThreadPool、SingleThreadPool
容许请求队列长度为Integer.MAX_VALUE可能会堆积大量请求从而致使OOM

CachedThreadPool、ScheduledThreadPool
容许建立线程数量为Integer.MAX_VALUE可能会建立大量线程从而致使OOM
复制代码

FixedThreadPool、SingleThreadPool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，因此可能会堆积大量请求从而致使OOM：markdown

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>()));
}
复制代码

CachedThreadPool、ScheduledThreadPool线程池maxSize使用Integer最大值，因此可能会建立大量线程从而致使OOM：架构

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());
}
复制代码

3 排查问题角度

你们使用Executors建立线程池应该经常使用以下语句：并发

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「JAVA前线」");
            }
        });
    }
}
复制代码

上述语句在功能层面是没有问题的，可是在生产环境中有可能遇到CPU飙高，线程数持续增长，内存溢出等问题，咱们时常须要经过线程快照进行观察。咱们经过jstack命令观察上述代码线程快照：ide

"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)
复制代码

咱们发现从线程快照看不出任何业务信息，只有相似pool-1-thread-2这种编号信息，不利于排查问题，咱们须要给线程命名。源码分析


4 为线程进行命名

在并发编程中咱们必定要为线程命名，这样有助于排查问题，关于如何命名咱们能够参考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：学习

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));
    }
}
复制代码

生产者默认线程名DubboServerHandler：测试

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

分析生产者线程快照发现DubboServerHandler：

"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)
复制代码

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";
}
复制代码

分析消费者线程快照发现DubboClientHandler：

"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)
复制代码

5 建立规范线程池

咱们参考DUBBO线程池定义命名工厂编写代码：

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;
    }
}
复制代码

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);
    }
}
复制代码

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「JAVA前线」");
            }
        });
        Thread.sleep(1000L);
    }
}
复制代码

再观察线程快照能够清晰查看业务名：

"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)
复制代码

6 文章总结

本文首先介绍了阿里巴巴JAVA开发手册不容许使用Executors建立线程池这个规定，而后从资源和排查问题两个角度分析了为何这么规定，最后咱们参考DUBBO线程池声明方式建立了一个规范线程池，这样使用线程池有助于快速定位和排查问题。

欢迎你们关注公众号「JAVA前线」查看更多精彩分享文章，主要包括源码分析、实际应用、架构思惟、职场分享、产品思考等等，同时欢迎你们加我微信「java_front」一块儿交流学习

相关文章
相关标签/搜索
每日一句
    每一个你不满意的现在，都有一个你没有努力的曾经。
本站公众号
   欢迎关注本站公众号,获取更多信息
联系我们 最近搜索 最新文章 沪ICP备13005482号-10 MyBatis教程 SQL 教程 MySQL教程 Java 教程 Thymeleaf 教程 Hibernate教程 Spring教程 Redis教程