优雅停机? 这个名词我是服的,如果抛开专业不谈,多好的名词啊!
其实优雅停机,就是在要关闭服务之前,不是立马全部关停,而是做好一些善后操作,比如:关闭线程、释放连接资源等。
再比如,就是不会让调用方的请求处理了一增,一下就中断了。而处理完本次后,再停止服务。
Java语言中,我们可以通过Runtime.getRuntime().addShutdownHook()方法来注册钩子,以保证程序平滑退出。(其他语言也类似)
来个栗子:
public class ShutdownGracefulTest { /** * 使用线程池处理任务 */ public static ExecutorService executorService = Executors.newCachedThreadPool(); public static void main(String[] args) { //假设有5个线程需要执行任务 for(int i = 0; i < 5; i++){ final int id = i; Thread taski = new Thread(new Runnable() { @Override public void run() { System.out.println(System.currentTimeMillis() + " : thread_" + id + " start..."); try { TimeUnit.SECONDS.sleep(id); } catch (InterruptedException e) { e.printStackTrace(); } System.out.println(System.currentTimeMillis() + " : thread_" + id + " finish!"); } }); taski.setDaemon(true); executorService.submit(taski); } Runtime.getRuntime().addShutdownHook(new Thread(new Runnable() { @Override public void run() { System.out.println(System.currentTimeMillis() + " : " + Thread.currentThread().getName() + " No1 shutdown hooking..."); boolean shutdown = true; try { executorService.shutdown(); System.out.println(System.currentTimeMillis() + " : " + Thread.currentThread().getName() + " shutdown signal got, wait threadPool finish."); executorService.awaitTermination(1500, TimeUnit.SECONDS); boolean done = false; System.out.println(System.currentTimeMillis() + " : " + Thread.currentThread().getName() + " all thread's done."); } catch (InterruptedException e) { e.printStackTrace(); // 尝试再次关闭 if(!executorService.isTerminated()) { executorService.shutdownNow(); } } System.out.println(System.currentTimeMillis() + " : " + Thread.currentThread().getName() + " No1 shutdown done..."); } })); Runtime.getRuntime().addShutdownHook(new Thread(new Runnable() { @Override public void run() { try { System.out.println(System.currentTimeMillis() + " : " + Thread.currentThread().getName() + " No2 shutdown hooking..."); Thread.sleep(1000); } catch (InterruptedException e) { e.printStackTrace(); } System.out.println(System.currentTimeMillis() + " : " + Thread.currentThread().getName() + " No2 shutdown done..."); } })); System.out.println("main method exit..."); System.exit(0); } }
运行结果如下:
很明显,确实是优雅了,虽然最后收到了一关闭信号,但是仍然保证了任务的处理完成。很棒吧!
那么,在实际应用中是如何体现优雅停机呢?
kill -15 pid
通过该命令发送一个关闭信号给到jvm, 然后就开始执行 Shutdown Hook 了,你可以做很多:
1. 关闭 socket 链接
2. 清理临时文件
3. 发送消息通知给订阅方,告知自己下线
4. 将自己将要被销毁的消息通知给子进程
5. 各种资源的释放
...
而在平时工作中,我们不乏看到很多运维同学,是这么干的:
kill -9 pid
如果这么干的话,jvm也无法了,kill -9 相当于一次系统宕机,系统断电。这会给应用杀了个措手不及,没有留给应用任何反应的机会。
所以,无论如何是优雅不起来了。
要优雅,是代码和运维的结合!
其中,线程池的关闭方式为:
executorService.shutdown(); executorService.awaitTermination(1500, TimeUnit.SECONDS);
ThreadPoolExecutor 在 shutdown 之后会变成 SHUTDOWN 状态,无法接受新的任务,随后等待正在执行的任务执行完成。意味着,shutdown 只是发出一个命令,至于有没有关闭还是得看线程自己。
ThreadPoolExecutor 对于 shutdownNow 的处理则不太一样,方法执行之后变成 STOP 状态,并对执行中的线程调用 Thread.interrupt() 方法(但如果线程未处理中断,则不会有任何事发生),所以并不代表“立刻关闭”。
shutdown() :启动顺序关闭,其中执行先前提交的任务,但不接受新任务。如果已经关闭,则调用没有附加效果。此方法不等待先前提交的任务完成执行。
shutdownNow():尝试停止所有正在执行的任务,停止等待任务的处理,并返回正在等待执行的任务的列表。当从此方法返回时,这些任务将从任务队列中耗尽(删除)。此方法不等待主动执行的任务终止。
executor.awaitTermination(this.awaitTerminationSeconds, TimeUnit.SECONDS)); 控制等待的时间,防止任务无限期的运行(前面已经强调过了,即使是 shutdownNow 也不能保证线程一定停止运行)。
注意:
虚拟机会对多个shutdownhook以未知的顺序调用,都执行完后再退出。
如果接收到 kill -15 pid 命令时,执行阻塞操作,可以做到等待任务执行完成之后再关闭 JVM。同时,也解释了一些应用执行 kill -15 pid 无法退出的问题,如:中断被阻塞了,或者hook运行了死循环代码。
实现原理:
Runtime.getRuntime().addShutdownHook(hook); // 添加钩子,开启优雅之路
// 具体流程如下:
/** * Registers a new virtual-machine shutdown hook. * * @param hook * An initialized but unstarted <tt>{@link Thread}</tt> object * * @throws IllegalArgumentException * If the specified hook has already been registered, * or if it can be determined that the hook is already running or * has already been run * * @throws IllegalStateException * If the virtual machine is already in the process * of shutting down * * @throws SecurityException * If a security manager is present and it denies * <tt>{@link RuntimePermission}("shutdownHooks")</tt> * * @see #removeShutdownHook * @see #halt(int) * @see #exit(int) * @since 1.3 */ public void addShutdownHook(Thread hook) { SecurityManager sm = System.getSecurityManager(); if (sm != null) { sm.checkPermission(new RuntimePermission("shutdownHooks")); } // 添加到 application 中 ApplicationShutdownHooks.add(hook); } // java.lang.ApplicationShutdownHooks.add(hook); static synchronized void add(Thread hook) { if(hooks == null) throw new IllegalStateException("Shutdown in progress"); if (hook.isAlive()) throw new IllegalArgumentException("Hook already running"); if (hooks.containsKey(hook)) throw new IllegalArgumentException("Hook previously registered"); // hooks 以map类型保存, k->k 形式存储,保证每一个钩子都是独立的 hooks.put(hook, hook); } // java.lang.ApplicationShutdownHooks 会先注册一个静态块,添加一个任务到 Shutdown 中 /* The set of registered hooks */ private static IdentityHashMap<Thread, Thread> hooks; static { try { Shutdown.add(1 /* shutdown hook invocation order */, false /* not registered if shutdown in progress */, new Runnable() { public void run() { // 即当该任务被调用时,调用自身的运行方法,使所有注册的 hook 运行起来 runHooks(); } } ); hooks = new IdentityHashMap<>(); } catch (IllegalStateException e) { // application shutdown hooks cannot be added if // shutdown is in progress. hooks = null; } } // runHooks 执行所有钩子线程,进行异步调用 /* Iterates over all application hooks creating a new thread for each * to run in. Hooks are run concurrently and this method waits for * them to finish. */ static void runHooks() { Collection<Thread> threads; synchronized(ApplicationShutdownHooks.class) { threads = hooks.keySet(); hooks = null; } for (Thread hook : threads) { hook.start(); } for (Thread hook : threads) { try { // 阻塞等待所有完成 hook.join(); } catch (InterruptedException x) { } } }
到现在为止,我们已经知道关闭钩子是如何执行的,但是,还不是知道,该钩子是何时触发?
// java.lang.Shutdown.add() 该方法会jvm主动调用,从而触发 后续钩子执行 /* Invoked by the JNI DestroyJavaVM procedure when the last non-daemon * thread has finished. Unlike the exit method, this method does not * actually halt the VM. */ static void shutdown() { synchronized (lock) { switch (state) { case RUNNING: /* Initiate shutdown */ state = HOOKS; break; case HOOKS: /* Stall and then return */ case FINALIZERS: break; } } synchronized (Shutdown.class) { // 执行序列 sequence(); } } // 而 sequence() 则会调用 runHooks(), 调用自定义的钩子任务 private static void sequence() { synchronized (lock) { /* Guard against the possibility of a daemon thread invoking exit * after DestroyJavaVM initiates the shutdown sequence */ if (state != HOOKS) return; } runHooks(); boolean rfoe; synchronized (lock) { state = FINALIZERS; rfoe = runFinalizersOnExit; } if (rfoe) runAllFinalizers(); } // 执行钩子,此处最多允许注册 10 个钩子,且进行同步调用,当然这是最顶级的钩子,钩子下还可以添加钩子,可以任意添加n个 private static void runHooks() { for (int i=0; i < MAX_SYSTEM_HOOKS; i++) { try { Runnable hook; synchronized (lock) { // acquire the lock to make sure the hook registered during // shutdown is visible here. currentRunningHook = i; hook = hooks[i]; } // 同步调用注册的hook, 即 前面看到 ApplicationShutdownHooks.runHooks() if (hook != null) hook.run(); } catch(Throwable t) { if (t instanceof ThreadDeath) { ThreadDeath td = (ThreadDeath)t; throw td; } } } }
如此,整个关闭流程完美了。
简化为:
1. 注册流程(应用主动调用)
Runtime.addShutdownHook -> ApplicationShutdownHooks.add()/static -> java.lang.Shutdown.add()/shutdown()
2. 执行流程(jvm自动调用)
java.lang.Shutdown.shutdown()->sequence()->runHooks() -> ApplicationShutdownHooks.runHooks() -> hooks 最终
不要害怕今日的苦,你要相信明天,更苦!