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OOMCachedThreadPool 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() {@Overridepublic 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 {@Overridepublic 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200int threads = url.getParameter(Constants.THREADS_KEY, Constants.DEFAULT_THREADS);// 队列容量默认0int 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);}@Overridepublic 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();}@Overridepublic 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() {@Overridepublic 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