Tomcat源码分析 （三）----- 生命周期机制 Lifecycle

tomcat里面有各种各样的组件，每个组件各司其职，组件之间又相互协作共同完成web服务器这样的工程。在这些组件之上，lifecycle（生命周期机制）至关重要！在学习各个组件之前，我们需要看看lifecycle是什么以及能做什么？实现原理又是怎样的？

什么是lifecycle?

lifecycle，其实就是一个状态机，对组件的由生到死状态的管理。

• 当组件在starting_prep、starting或started时，调用start()方法没有任何效果
• 当组件在new状态时，调用start()方法会导致init()方法被立刻执行，随后start()方法被执行
• 当组件在stopping_prep、stopping或stopped时，调用stop()方法没有任何效果
• 当一个组件在new状态时，调用stop()方法会将组件状态变更为stopped，比较典型的场景就是组件启动失败，其子组件还没有启动。当一个组件停止的时候，它将尝试停止它下面的所有子组件，即使子组件还没有启动。

lifecycle方法

我们看看lifecycle有哪些方法，如下所示：

public interface lifecycle {
    // 添加监听器
    public void addlifecyclelistener(lifecyclelistener listener);
    // 获取所以监听器
    public lifecyclelistener[] findlifecyclelisteners();
    // 移除某个监听器
    public void removelifecyclelistener(lifecyclelistener listener);
    // 初始化方法
    public void init() throws lifecycleexception;
    // 启动方法
    public void start() throws lifecycleexception;
    // 停止方法，和start对应
    public void stop() throws lifecycleexception;
    // 销毁方法，和init对应
    public void destroy() throws lifecycleexception;
    // 获取生命周期状态
    public lifecyclestate getstate();
    // 获取字符串类型的生命周期状态
    public string getstatename();
}

lifecyclebase

lifecyclebase是lifecycle的基本实现。我们逐一来看lifecycle的各个方法。

增加、删除和获取监听器

private final list<lifecyclelistener> lifecyclelisteners = new copyonwritearraylist<>();

@override
public void addlifecyclelistener(lifecyclelistener listener) {
    lifecyclelisteners.add(listener);
}
@override
public lifecyclelistener[] findlifecyclelisteners() {
    return lifecyclelisteners.toarray(new lifecyclelistener[0]);
}
@override
public void removelifecyclelistener(lifecyclelistener listener) {
    lifecyclelisteners.remove(listener);
}

1. 生命周期监听器保存在一个线程安全的list中，copyonwritearraylist。所以add和remove都是直接调用此list的相应方法。
2. findlifecyclelisteners返回的是一个数组，为了线程安全，所以这儿会生成一个新数组。

init()

@override
public final synchronized void init() throws lifecycleexception {
    // 非new状态，不允许调用init()方法
    if (!state.equals(lifecyclestate.new)) {
        invalidtransition(lifecycle.before_init_event);
    }

    try {
        // 初始化逻辑之前，先将状态变更为`initializing`
        setstateinternal(lifecyclestate.initializing, null, false);
        // 初始化，该方法为一个abstract方法，需要组件自行实现
        initinternal();
        // 初始化完成之后，状态变更为`initialized`
        setstateinternal(lifecyclestate.initialized, null, false);
    } catch (throwable t) {
        // 初始化的过程中，可能会有异常抛出，这时需要捕获异常，并将状态变更为`failed`
        exceptionutils.handlethrowable(t);
        setstateinternal(lifecyclestate.failed, null, false);
        throw new lifecycleexception(
                sm.getstring("lifecyclebase.initfail",tostring()), t);
    }
}

setstateinternal方法用于维护状态，同时在状态转换成功之后触发事件。为了状态的可见性，所以state声明为volatile类型的。

private volatile lifecyclestate state = lifecyclestate.new;。

private synchronized void setstateinternal(lifecyclestate state,
        object data, boolean check) throws lifecycleexception {
    if (log.isdebugenabled()) {
        log.debug(sm.getstring("lifecyclebase.setstate", this, state));
    }

    // 是否校验状态
    if (check) {
        // must have been triggered by one of the abstract methods (assume
        // code in this class is correct)
        // null is never a valid state
        // state不允许为null
        if (state == null) {
            invalidtransition("null");
            // unreachable code - here to stop eclipse complaining about
            // a possible npe further down the method
            return;
        }

        // any method can transition to failed
        // startinternal() permits starting_prep to starting
        // stopinternal() permits stopping_prep to stopping and failed to
        // stopping
        if (!(state == lifecyclestate.failed ||
                (this.state == lifecyclestate.starting_prep &&
                        state == lifecyclestate.starting) ||
                (this.state == lifecyclestate.stopping_prep &&
                        state == lifecyclestate.stopping) ||
                (this.state == lifecyclestate.failed &&
                        state == lifecyclestate.stopping))) {
            // no other transition permitted
            invalidtransition(state.name());
        }
    }

    // 设置状态
    this.state = state;
    // 触发事件
    string lifecycleevent = state.getlifecycleevent();
    if (lifecycleevent != null) {
        firelifecycleevent(lifecycleevent, data);
    }
}

我们看看firelifecycleevent方法，

public void firelifecycleevent(string type, object data) {
    // 事件监听,观察者模式的另一种方式
    lifecycleevent event = new lifecycleevent(lifecycle, type, data);
    lifecyclelistener interested[] = listeners;// 监听器数组 关注 事件(启动或者关闭事件)
    // 循环通知所有生命周期时间侦听器
    for (int i = 0; i < interested.length; i++)
        // 每个监听器都有自己的逻辑
        interested[i].lifecycleevent(event);
}

 首先, 创建一个事件对象, 然通知所有的监听器发生了该事件.并做响应.

start()

@override
public final synchronized void start() throws lifecycleexception {
    // `starting_prep`、`starting`和`started时，将忽略start()逻辑
    if (lifecyclestate.starting_prep.equals(state) || lifecyclestate.starting.equals(state) ||
            lifecyclestate.started.equals(state)) {

        if (log.isdebugenabled()) {
            exception e = new lifecycleexception();
            log.debug(sm.getstring("lifecyclebase.alreadystarted", tostring()), e);
        } else if (log.isinfoenabled()) {
            log.info(sm.getstring("lifecyclebase.alreadystarted", tostring()));
        }

        return;
    }

    // `new`状态时，执行init()方法
    if (state.equals(lifecyclestate.new)) {
        init();
    }

    // `failed`状态时，执行stop()方法
    else if (state.equals(lifecyclestate.failed)) {
        stop();
    }

    // 不是`initialized`和`stopped`时，则说明是非法的操作
    else if (!state.equals(lifecyclestate.initialized) &&
            !state.equals(lifecyclestate.stopped)) {
        invalidtransition(lifecycle.before_start_event);
    }

    try {
        // start前的状态设置
        setstateinternal(lifecyclestate.starting_prep, null, false);
        // start逻辑，抽象方法，由组件自行实现
        startinternal();
        // start过程中，可能因为某些原因失败，这时需要stop操作
        if (state.equals(lifecyclestate.failed)) {
            // this is a 'controlled' failure. the component put itself into the
            // failed state so call stop() to complete the clean-up.
            stop();
        } else if (!state.equals(lifecyclestate.starting)) {
            // shouldn't be necessary but acts as a check that sub-classes are
            // doing what they are supposed to.
            invalidtransition(lifecycle.after_start_event);
        } else {
            // 设置状态为started
            setstateinternal(lifecyclestate.started, null, false);
        }
    } catch (throwable t) {
        // this is an 'uncontrolled' failure so put the component into the
        // failed state and throw an exception.
        exceptionutils.handlethrowable(t);
        setstateinternal(lifecyclestate.failed, null, false);
        throw new lifecycleexception(sm.getstring("lifecyclebase.startfail", tostring()), t);
    }
}

stop()

@override
public final synchronized void stop() throws lifecycleexception {
    // `stopping_prep`、`stopping`和stopped时，将忽略stop()的执行
    if (lifecyclestate.stopping_prep.equals(state) || lifecyclestate.stopping.equals(state) ||
            lifecyclestate.stopped.equals(state)) {

        if (log.isdebugenabled()) {
            exception e = new lifecycleexception();
            log.debug(sm.getstring("lifecyclebase.alreadystopped", tostring()), e);
        } else if (log.isinfoenabled()) {
            log.info(sm.getstring("lifecyclebase.alreadystopped", tostring()));
        }

        return;
    }

    // `new`状态时，直接将状态变更为`stopped`
    if (state.equals(lifecyclestate.new)) {
        state = lifecyclestate.stopped;
        return;
    }

    // stop()的执行，必须要是`started`和`failed`
    if (!state.equals(lifecyclestate.started) && !state.equals(lifecyclestate.failed)) {
        invalidtransition(lifecycle.before_stop_event);
    }

    try {
        // `failed`时，直接触发before_stop_event事件
        if (state.equals(lifecyclestate.failed)) {
            // don't transition to stopping_prep as that would briefly mark the
            // component as available but do ensure the before_stop_event is
            // fired
            firelifecycleevent(before_stop_event, null);
        } else {
            // 设置状态为stopping_prep
            setstateinternal(lifecyclestate.stopping_prep, null, false);
        }

        // stop逻辑，抽象方法，组件自行实现
        stopinternal();

        // shouldn't be necessary but acts as a check that sub-classes are
        // doing what they are supposed to.
        if (!state.equals(lifecyclestate.stopping) && !state.equals(lifecyclestate.failed)) {
            invalidtransition(lifecycle.after_stop_event);
        }
        // 设置状态为stopped
        setstateinternal(lifecyclestate.stopped, null, false);
    } catch (throwable t) {
        exceptionutils.handlethrowable(t);
        setstateinternal(lifecyclestate.failed, null, false);
        throw new lifecycleexception(sm.getstring("lifecyclebase.stopfail",tostring()), t);
    } finally {
        if (this instanceof lifecycle.singleuse) {
            // complete stop process first
            setstateinternal(lifecyclestate.stopped, null, false);
            destroy();
        }
    }
}

destroy()

@override
public final synchronized void destroy() throws lifecycleexception {
    // `failed`状态时，直接触发stop()逻辑
    if (lifecyclestate.failed.equals(state)) {
        try {
            // triggers clean-up
            stop();
        } catch (lifecycleexception e) {
            // just log. still want to destroy.
            log.warn(sm.getstring(
                    "lifecyclebase.destroystopfail", tostring()), e);
        }
    }

    // `destroying`和`destroyed`时，忽略destroy的执行
    if (lifecyclestate.destroying.equals(state) ||
            lifecyclestate.destroyed.equals(state)) {

        if (log.isdebugenabled()) {
            exception e = new lifecycleexception();
            log.debug(sm.getstring("lifecyclebase.alreadydestroyed", tostring()), e);
        } else if (log.isinfoenabled() && !(this instanceof lifecycle.singleuse)) {
            // rather than have every component that might need to call
            // destroy() check for singleuse, don't log an info message if
            // multiple calls are made to destroy()
            log.info(sm.getstring("lifecyclebase.alreadydestroyed", tostring()));
        }

        return;
    }

    // 非法状态判断
    if (!state.equals(lifecyclestate.stopped) &&
            !state.equals(lifecyclestate.failed) &&
            !state.equals(lifecyclestate.new) &&
            !state.equals(lifecyclestate.initialized)) {
        invalidtransition(lifecycle.before_destroy_event);
    }

    try {
        // destroy前状态设置
        setstateinternal(lifecyclestate.destroying, null, false);
       // 抽象方法，组件自行实现
        destroyinternal();
        // destroy后状态设置
        setstateinternal(lifecyclestate.destroyed, null, false);
    } catch (throwable t) {
        exceptionutils.handlethrowable(t);
        setstateinternal(lifecyclestate.failed, null, false);
        throw new lifecycleexception(
                sm.getstring("lifecyclebase.destroyfail",tostring()), t);
    }
}

模板方法

从上述源码看得出来，lifecyclebase是使用了状态机+模板模式来实现的。模板方法有下面这几个：

// 初始化方法
protected abstract void initinternal() throws lifecycleexception;
// 启动方法
protected abstract void startinternal() throws lifecycleexception;
// 停止方法
protected abstract void stopinternal() throws lifecycleexception;
// 销毁方法
protected abstract void destroyinternal() throws lifecycleexception;

总结

lifecycle其实非常简单，代码也不复杂，但是剖析其实现对于我们理解组件的生命周期有很大的帮助，也有助于我们对设计模式的回顾。