【问题记录】线程池死锁问题

今天看到个父子线程提交导致线程池死锁的问题，把该问题复盘一下

模拟

    public static void main(String[] args) throws Exception {
        ThreadPoolExecutor executor = new ThreadPoolExecutor(10, 10, 60, TimeUnit.SECONDS, new LinkedBlockingQueue<>());
        for (int i = 0; i < 10; i++) {
            executor.execute(() -> {
                try {
                    System.out.println("父任务开始");
                    TimeUnit.SECONDS.sleep(5);
                    Future<?> future = executor.submit(() -> System.out.println("子任务开始"));
                    System.out.println("等待子任务结果:" + future.get());
                } catch (InterruptedException e) {
                    e.printStackTrace();
                } catch (ExecutionException e) {
                    throw new RuntimeException(e);
                }
            });
        }
        TimeUnit.SECONDS.sleep(6);
        System.out.println(executor.getQueue());

//        executor.shutdown();

    }

分析

首先看下线程池配置，10 个核心线程数，10 个最大线程数，无界的任务队列。开始一下子提交了 10 个父任务，于是核心线程立刻被用完了，后面的任务只能进阻塞队列。每个父任务等待 5 秒（模拟业务耗时）后开始提交子任务，并同步等子任务的结果，然后死锁就产生了

因为所有父任务的子任务都在阻塞队列中，他们在等执行父任务的工作线程空闲后把他们取出来执行；而所有父任务都在等子任务完成，这样就产生了死锁

在 main 线程 6 秒后，我打印了任务队列的情况，确实有 10 个子任务。再 jstack 确认下，这里截取了一部分

"pool-1-thread-4" #14 prio=5 os_prio=31 tid=0x00007f95a696c000 nid=0x5b03 waiting on condition [0x00007000054cb000]
   java.lang.Thread.State: WAITING (parking)
	at sun.misc.Unsafe.park(Native Method)
	- parking to wait for  <0x000000076bbb0780> (a java.util.concurrent.FutureTask)
	at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
	at java.util.concurrent.FutureTask.awaitDone(FutureTask.java:429)
	at java.util.concurrent.FutureTask.get(FutureTask.java:191)
	at com.cyk.lottery.application.mq.Main.lambda$main$1(Main.java:22)
	at com.cyk.lottery.application.mq.Main$$Lambda$1/1717159510.run(Unknown Source)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	at java.lang.Thread.run(Thread.java:748)

"pool-1-thread-3" #13 prio=5 os_prio=31 tid=0x00007f95a60df800 nid=0x7b03 waiting on condition [0x00007000053c8000]
   java.lang.Thread.State: WAITING (parking)
	at sun.misc.Unsafe.park(Native Method)
	- parking to wait for  <0x000000076b68dd58> (a java.util.concurrent.FutureTask)
	at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
	at java.util.concurrent.FutureTask.awaitDone(FutureTask.java:429)
	at java.util.concurrent.FutureTask.get(FutureTask.java:191)
	at com.cyk.lottery.application.mq.Main.lambda$main$1(Main.java:22)
	at com.cyk.lottery.application.mq.Main$$Lambda$1/1717159510.run(Unknown Source)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	at java.lang.Thread.run(Thread.java:748)

"pool-1-thread-2" #12 prio=5 os_prio=31 tid=0x00007f95a60de800 nid=0x5903 waiting on condition [0x00007000052c5000]
   java.lang.Thread.State: WAITING (parking)
	at sun.misc.Unsafe.park(Native Method)
	- parking to wait for  <0x000000076ba69098> (a java.util.concurrent.FutureTask)
	at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)

也可以看到所有父任务都处于休眠状态，卡在 FutureTask#get

解决方案

如果生产发生这样的问题，无疑是一场灾难，线程池会因此直接瘫痪，并且如果后续的任务不断涌进任务队列，也有让应用 OOM 的风险

怎么解决呢？其实解决方案很简单，就是区分父子任务的线程池，不要混用 

参考

线程池运用不当的一次线上事故 | HeapDump性能社区