多线程死锁实例与定位

既然可以上锁，那么假如有2个线程，一个线程想先锁对象1，再锁对象2，恰好另外有一个线程先锁对象2，再锁对象1。在这个过程中，当线程1把对象1锁好以后，就想去锁对象2，但是不巧，线程2已经把对象2锁上了，也正在尝试去锁对象1。什么时候结束呢，只有线程1把2个对象都锁上并把方法执行完，并且线程2把2个对象也都锁上并且把方法执行完毕，那么就结束了，但是，谁都不肯放掉已经锁上的对象，所以就没有结果，这种情况就叫做线程死锁。

死锁实例

public class DeadThread implements Runnable{

    private  static Object object1 = new Object();
    private  static Object object2 = new Object();

    private boolean flag;

    public DeadThread(boolean flag){
        this.flag=flag;
    }

    @Override
    public void run() {
        if(flag){
            synchronized(object1){
                System.out.println(Thread.currentThread().getName()+"--1");
                try {
                    Thread.sleep(200);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
                synchronized(object2){
                    System.out.println(Thread.currentThread().getName()+"--2");
                }
            }
        }else{
            synchronized(object2){
                System.out.println(Thread.currentThread().getName()+"--2");
                try {
                    Thread.sleep(200);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
                synchronized(object1){
                    System.out.println(Thread.currentThread().getName()+"--1");
                }
            }
        }
    }


    public static void main(String[] args) {
        DeadThread thread1=new DeadThread(true);
        DeadThread thread2=new DeadThread(false);
        new Thread(thread1).start();
        new Thread(thread2).start();
    }
}

程序死锁定位

1 获取java进程
bogon:effectivejava mazh$ jps
9521 Launcher
2466 
9523 DeadThread --你的java类名
95269 Launcher
9543 Jps
67211 Launcher
4398 RemoteMavenServer

此处找到java进程pid为9523

2 使用jstack指令获取锁详细栈信息

bogon:effectivejava mazh$ jstack 9523
2019-10-11 18:34:08
Full thread dump Java HotSpot(TM) 64-Bit Server VM (25.181-b13 mixed mode):

"Attach Listener" #13 daemon prio=9 os_prio=31 tid=0x00007fb951880800 nid=0xc07 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"DestroyJavaVM" #12 prio=5 os_prio=31 tid=0x00007fb952856000 nid=0x2603 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"Thread-1" #11 prio=5 os_prio=31 tid=0x00007fb95182c800 nid=0x3f03 waiting for monitor entry [0x0000700003b8b000]
   java.lang.Thread.State: BLOCKED (on object monitor)
        at concurrency.deathlock.Deathlock.run(Deathlock.java:42)
        - waiting to lock <0x000000076ab0b958> (a java.lang.String)
        - locked <0x000000076ab0bde0> (a java.lang.String)
        at java.lang.Thread.run(Thread.java:748)

"Thread-0" #10 prio=5 os_prio=31 tid=0x00007fb95182b800 nid=0x3d03 waiting for monitor entry [0x0000700003a88000]
   java.lang.Thread.State: BLOCKED (on object monitor)
        at concurrency.deathlock.Deathlock.run(Deathlock.java:30)
        - waiting to lock <0x000000076ab0bde0> (a java.lang.String)
        - locked <0x000000076ab0b958> (a java.lang.String)
        at java.lang.Thread.run(Thread.java:748)
....................
....................(省略无效内容)
Found ones Java-level deadlock:
=============================
"Thread-1":
  waiting to lock monitor 0x00007fb95200c158 (object 0x000000076ab0b958, a java.lang.String),
  which is held by "Thread-0"
"Thread-0":
  waiting to lock monitor 0x00007fb952009818 (object 0x000000076ab0bde0, a java.lang.String),
  which is held by "Thread-1"

Java stack information for the threads listed above:
===================================================
"Thread-1":
        at concurrency.deathlock.Deathlock.run(Deathlock.java:42)
        - waiting to lock <0x000000076ab0b958> (a java.lang.String)
        - locked <0x000000076ab0bde0> (a java.lang.String)
        at java.lang.Thread.run(Thread.java:748)
"Thread-0":
        at concurrency.deathlock.Deathlock.run(Deathlock.java:30)
        - waiting to lock <0x000000076ab0bde0> (a java.lang.String)
        - locked <0x000000076ab0b958> (a java.lang.String)
        at java.lang.Thread.run(Thread.java:748)

Found 1 deadlock.

重点部分如下：
===================================================
"Thread-1":
        at concurrency.deathlock.Deathlock.run(Deathlock.java:42)
        - waiting to lock <0x000000076ab0b958> (a java.lang.String)
        - locked <0x000000076ab0bde0> (a java.lang.String)
        at java.lang.Thread.run(Thread.java:748)
"Thread-0":
        at concurrency.deathlock.Deathlock.run(Deathlock.java:30)
        - waiting to lock <0x000000076ab0bde0> (a java.lang.String)
        - locked <0x000000076ab0b958> (a java.lang.String)
        at java.lang.Thread.run(Thread.java:748)

Found 1 deadlock.
告诉你死锁的代码位置了

死循环导致的死锁排查方法
可以通过linux下top命令查看cpu使用率较高的java进程，进而用top -Hp ➕pid查看该java进程下cpu使用率较高的线程。再用jstack命令查看线程具体调用情况，排查问题。

避免死锁的方法

1 不要多层级加锁，如不要先获取锁1再获取锁2，尽量使用一个锁
2 相同顺序加锁，所有涉及到锁1和锁2的，都是先锁1再锁2
3 可以使用lock.tryLock(timeout, unit)等带有超时时间的方法上锁，超时会自动放弃 （mysql中有超时配置和死锁检测）

怎么查看线程的状态？

1)thread.getState();可以打印线程状态
2)jps获取到线程， jstack pid 获取线程状态
3)使用jconsole图形化工具

死锁的时候线程是什么状态？拿不到锁，所以在在锁等待池当中
子线程future.get，主线程什么状态？
get本质使用了AQS架构，调用await将主线程放到等待队列中。