Java多线程之间引发死锁问题(1)

死锁代码：

import java.util.concurrent.CountDownLatch;
 
public class DeadLock {
    
    public static String obj1 = "obj1";//资源1
    public static String obj2 = "obj2";//资源2
    public static void main(String[] args) throws Exception{
        final CountDownLatch countdown = new CountDownLatch(1);
        new Thread(new Runnable() {
            @Override
            public void run() {
                try {
                    countdown.await();//都进入等待
                    synchronized(DeadLock.obj1){
                        System.out.println(Thread.currentThread()+"has lock obj1");
                        Thread.sleep(3000);//获取obj1后先等一会儿，让Lock2有足够的时间锁住obj2
                        System.out.println(Thread.currentThread()+"ready to lock obj2");
                        synchronized(DeadLock.obj2){
                            System.out.println(Thread.currentThread()+"has lock obj2");
                        }
                    }
                }catch (Exception e) {
                    e.printStackTrace();
                }
            }
        },"thread1" ).start();
 
        new Thread(new Runnable() {
            @Override
            public void run() {
                try {
                    countdown.await();//都进入等待
                    synchronized(DeadLock.obj2){
                        System.out.println(Thread.currentThread()+"has lock obj2");
                        Thread.sleep(3000);//获取obj1后先等一会儿，让Lock2有足够的时间锁住obj2
                        System.out.println(Thread.currentThread()+"ready to lock obj1");
                        synchronized(DeadLock.obj1){
                            System.out.println(Thread.currentThread()+"has lock obj1");
                        }
                    }
                }catch (Exception e) {
                    e.printStackTrace();
                }
            }
        },"thread2" ).start();
 
        countdown.countDown();
    }
 
}

如何在命令行查看死锁问题？

①jps -l 查看java进程信息；

②jstack -l pid | more  查看当前java进程的堆栈状态

可以看到两个线程之间互相占用对方资源，等待释放，产生阻塞死锁。

在top命令中，已经获取到了占用cpu资源较高的线程pid，将该pid转成16进制的值，在thread dump中每个线程都有一个nid，找到对应的nid即可；隔段时间再执行一次stack命令获取thread dump，区分两份dump是否有差别，在nid=0x229c的线程调用栈中，发现该线程一直在执行DeadLock类第40行的run方法，得到这个信息，就可以检查对应的代码是否有问题。

解决方法： https://blog.csdn.net/MrLiar17/article/details/86623614