tomcat8.5.57源码阅读笔记4

Bootstrap#start()

daemon = bootstrap, 所以调用的还是 org.apache.catalina.startup.Bootstrap#start()

public void start() throws Exception {
    if (catalinaDaemon == null) {
        init();
    }

    Method method = catalinaDaemon.getClass().getMethod("start", (Class[]) null);
    // 反射调用 Catalina#start()
    method.invoke(catalinaDaemon, (Object[]) null);
}

反射调用了 Catalina#start() 方法, 进行启动工作.

Catalina#start()

public void start() {
    if (getServer() == null) {
        load();
    }
    if (getServer() == null) {
        log.fatal("Cannot start server. Server instance is not configured.");
        return;
    }

    //获取当前纳秒值
    long t1 = System.nanoTime();

    // Start the new server
    try {
        //StandardServer#start(), 最终调用的是 StandardServer#startInternal()
        getServer().start();
    } catch (LifecycleException e) {
        log.fatal(sm.getString("catalina.serverStartFail"), e);
        try {
            getServer().destroy();
        } catch (LifecycleException e1) {
            log.debug("destroy() failed for failed Server ", e1);
        }
        return;
    }

    //再次获取当前纳秒值, 计算启动花了多长时间
    long t2 = System.nanoTime();
    if(log.isInfoEnabled()) {
        log.info("Server startup in " + ((t2 - t1) / 1000000) + " ms");
    }

    // Register shutdown hook
    // 注册回调方法, 用于安全关闭服务
    if (useShutdownHook) {
        if (shutdownHook == null) {
            shutdownHook = new CatalinaShutdownHook();
        }
        Runtime.getRuntime().addShutdownHook(shutdownHook);

        // If JULI is being used, disable JULI's shutdown hook since
        // shutdown hooks run in parallel and log messages may be lost
        // if JULI's hook completes before the CatalinaShutdownHook()
        LogManager logManager = LogManager.getLogManager();
        if (logManager instanceof ClassLoaderLogManager) {
            ((ClassLoaderLogManager) logManager).setUseShutdownHook(
                    false);
        }
    }

    // Bootstrap中会设置await为true，其目的在于让tomcat在shutdown端口阻塞监听关闭命令
    if (await) {
        await();
        stop();
    }
}

这里继续调用了 Server 的 start 方法. 实际调用的是 org.apache.catalina.util.LifecycleBase#start(),

然后此方法中调用了抽象方法 startInternal(), 也就是 StandardServer#startInternal()

StandardServer#startInternal()

protected void startInternal() throws LifecycleException {
    // 事件通知 CONFIGURE_START_EVENT -> configure_start
    fireLifecycleEvent(CONFIGURE_START_EVENT, null);
    // 修改tomcat状态为 STARTING, 这一步也会进行事件通知 START_EVENT -> start
    setState(LifecycleState.STARTING);

    //NamingResourcesImpl#startInternal()   全局资源
    //在server.xml中配置了该变量, pathname="conf/tomcat-users.xml"
    globalNamingResources.start();

    // Start our defined Services
    synchronized (servicesLock) {
        for (Service service : services) {
            // StandardService.startInternal()
            service.start();
        }
    }
}

service.start() 最终调用的是 StandardService.startInternal()

StandardService.startInternal()

protected void startInternal() throws LifecycleException {

    if(log.isInfoEnabled())
        log.info(sm.getString("standardService.start.name", this.name));
    setState(LifecycleState.STARTING);

    // Start our defined Container first
    if (engine != null) {
        synchronized (engine) {
            //StandardEngine#start() 最终调用 StandardEngine#startInternal()
            //这一句代码, 就将 Host, Context, Wrapper 的 start 执行到了
            //host是通过 server.xml 解析得到
            //context是通过扫描文件的方式解析得到, 有三种方式: xml配置方式, war包方式, 文件夹方式
            //wrapper也是通过扫描jar包的方式得到
            engine.start();
        }
    }

    // 启动Executor线程池, 默认情况下, 是空, 通过修改 server.xml中的tomcatThreadPool来改变
    synchronized (executors) {
        for (Executor executor: executors) {
            executor.start();
        }
    }

    // 启动 MapperListener, 进行注册功能
    // 最终会调用 MapperListener#startInternal() 方法
    //* 这个方法主要干了两件事情
    //* 1. 将 MapperListener 注册到容器和子容器(Host, Context, Wrapper)的 listeners 和 lifecycleListeners 中
    //* 2. 注册 host, context, wrapper
    mapperListener.start();

    // Start our defined Connectors second
    synchronized (connectorsLock) {
        for (Connector connector: connectors) {
            try {
                // If it has already failed, don't try and start it
                if (connector.getState() != LifecycleState.FAILED) {
                    //最终调用 Connector.startInternal()
                    connector.start();
                }
            } catch (Exception e) {
                log.error(sm.getString(
                        "standardService.connector.startFailed",
                        connector), e);
            }
        }
    }
}

这里和初始化的时候很像, 初始化的时候, 是对他们分别初始化, 而这里, 是对他们分别启动

1. engine.start() : 启动 engine, 里面会启动 Host, Context, Wrapper. 最终会调用 StandardEngine#startInternal() 

2. executor.start() : 启动线程池, 默认情况下, 这个线程池为空.

3. mapperListener.start() : 将MapperListener注册到 Host,Context,Wrapper 的监听器中, 然后对 Host, Context, Wrapper 进行映射

4. connector.start() : 启动连接器

StandardEngine#startInternal() 

  

protected synchronized void startInternal() throws LifecycleException {

    // Log our server identification information
    if(log.isInfoEnabled())
        log.info( "Starting Servlet Engine: " + ServerInfo.getServerInfo());

    // Standard container startup
    //通过解析 server.xml 文件, 可以拿到 engine下面有一个 host, 也就是说, StandardEngine.children有一个host
    super.startInternal();
}

super.startInternal() 调用的是父类 org.apache.catalina.core.ContainerBase#startInternal() 方法:

protected synchronized void startInternal() throws LifecycleException {

    // Start our subordinate components, if any
    logger = null;
    getLogger();
    //集群客户端
    Cluster cluster = getClusterInternal();
    if (cluster instanceof Lifecycle) {
        //如果有集群的话, 则会启动集群
        ((Lifecycle) cluster).start();
    }
    //域
    Realm realm = getRealmInternal();
    if (realm instanceof Lifecycle) {
        ((Lifecycle) realm).start();
    }

    // Start our child containers, if any
    // 把子容器的启动步骤放在线程中处理，默认情况下线程池只有一个线程处理任务队列
    //StandardEngine 调用的时候, 这里拿到的 children 有一个值: StandardHost[localhost]
    // <Engine name="Catalina" defaultHost="localhost">
    //      <Host name="localhost"  appBase="webapps" unpackWARs="true" autoDeploy="true" startStopThreads="1">
    // </engine>
    Container children[] = findChildren();
    List<Future<Void>> results = new ArrayList<>();
    for (Container child : children) {
        //StartChild 的 call方法, 调用的是 child.start() 方法, 最终会调用 StandardHost.startInternal
        results.add(startStopExecutor.submit(new StartChild(child)));
    }

    MultiThrowable multiThrowable = null;

    // 阻塞当前线程，直到子容器start完成
    for (Future<Void> result : results) {
        try {
            result.get();
        } catch (Throwable e) {
            log.error(sm.getString("containerBase.threadedStartFailed"), e);
            if (multiThrowable == null) {
                multiThrowable = new MultiThrowable();
            }
            multiThrowable.add(e);
        }

    }
    if (multiThrowable != null) {
        throw new LifecycleException(sm.getString("containerBase.threadedStartFailed"),
                multiThrowable.getThrowable());
    }

    // Start the Valves in our pipeline (including the basic), if any
    // 启用 Pipeline --> StandardPipeline#startInternal()
    if (pipeline instanceof Lifecycle) {
        ((Lifecycle) pipeline).start();
    }

    //StandardHost调用时, 激发 STARTING 监听器 HostConfig,最终会调用 HostConfig#start() 方法 - 这一步很关键
    setState(LifecycleState.STARTING);

    // Start our thread
    // 开启ContainerBackgroundProcessor线程用于调用子容器的backgroundProcess方法，
    // 默认情况下backgroundProcessorDelay=-1，不会启用该线程
    threadStart();
}

1. 通过Future框架异步调用 StandardHost#startInternal() 方法, 并阻塞等待所有当前 engine 下所有的 Host 启动完成

2. setState() 时, 会激发监听器 HostConfig. 这个HostConfig 是在 解析 Server.xml 的时候, 创建并绑定的.

StandardHost#startInternal()

protected synchronized void startInternal() throws LifecycleException {
    // Set error report valve
    // errorValve默认使用 ErrorReportValve
    String errorValve = getErrorReportValveClass();
    if ((errorValve != null) && (!errorValve.equals(""))) {
        try {
            boolean found = false;
            // 如果所有的阀门中已经存在这个实例，则不进行处理，否则添加到 Pipeline 中
            Valve[] valves = getPipeline().getValves();
            for (Valve valve : valves) {
                if (errorValve.equals(valve.getClass().getName())) {
                    found = true;
                    break;
                }
            }
            // 如果未找到则添加到 Pipeline 中，注意是添加到 basic valve 的前面
            // 默认情况下，first valve 是 AccessLogValve，basic 是 StandardHostValve
            if(!found) {
                Valve valve =
                    (Valve) Class.forName(errorValve).getConstructor().newInstance();
                getPipeline().addValve(valve);
            }
        } catch (Throwable t) {
            ExceptionUtils.handleThrowable(t);
            log.error(sm.getString(
                    "standardHost.invalidErrorReportValveClass",
                    errorValve), t);
        }
    }
    // 调用父类 ContainerBase，完成统一的启动动作
    super.startInternal();
}

这里的 startInternal() 就是上面的 org.apache.catalina.core.ContainerBase#startInternal() 方法.

但是此时, host 的 children 是空的, 所以里面并没有能够调用 StandardContext#startInternal()

HostConfig#lifecycleEvent()

HostConfig 实现了 LifecycleListener 接口. 所以实际上, 他是一个监听器. 

setState() 的时候, 会激发 HostConfig的监听器方法 lifecycleEvent(), 

public void lifecycleEvent(LifecycleEvent event) {

    // Identify the host we are associated with
    try {
        host = (Host) event.getLifecycle();
        if (host instanceof StandardHost) {
            setCopyXML(((StandardHost) host).isCopyXML());
            setDeployXML(((StandardHost) host).isDeployXML());
            setUnpackWARs(((StandardHost) host).isUnpackWARs());
            setContextClass(((StandardHost) host).getContextClass());
        }
    } catch (ClassCastException e) {
        log.error(sm.getString("hostConfig.cce", event.getLifecycle()), e);
        return;
    }

    // Process the event that has occurred
    //判断事件是否由 Host 发出，并且为 HostConfig 设置属性
    if (event.getType().equals(Lifecycle.PERIODIC_EVENT)) {
        check();
    } else if (event.getType().equals(Lifecycle.BEFORE_START_EVENT)) {
        beforeStart();
    } else if (event.getType().equals(Lifecycle.START_EVENT)) {
        start();
    } else if (event.getType().equals(Lifecycle.STOP_EVENT)) {
        stop();
    }
}

根据当前事件的状态, 会调用这里的 start() 方法. 此start 方法中, 会执行 部署 webapp 的方法 : org.apache.catalina.startup.HostConfig#deployApps()

protected void deployApps() {
    File appBase = host.getAppBaseFile();
    File configBase = host.getConfigBaseFile();
    // 过滤出 webapp 要部署应用的目录
    String[] filteredAppPaths = filterAppPaths(appBase.list());
    // Deploy XML descriptors from configBase
    // 1. xml部署 - 不推荐这么使用
    // <host><context docBase="D://abc/eee" path="sdm" reloadable="true"></context></host>
    // 部署 xml 描述文件
    deployDescriptors(configBase, configBase.list());
    // Deploy WARs
    // 2. war包部署
    // 解压 war 包，但是这里还不会去启动应用
    deployWARs(appBase, filteredAppPaths);
    // Deploy expanded folders
    // 3. 目录部署
    // 处理已经存在的目录，前面解压的 war 包不会再行处理
    deployDirectories(appBase, filteredAppPaths);
}

源码里面放的几个 Context , 都是目录结构的, 所以会走 deployDirectories() 方法

protected void deployDirectories(File appBase, String[] files) {

    if (files == null)
        return;

    ExecutorService es = host.getStartStopExecutor();
    List<Future<?>> results = new ArrayList<>();

    for (String file : files) {

        if (file.equalsIgnoreCase("META-INF"))
            continue;
        if (file.equalsIgnoreCase("WEB-INF"))
            continue;
        File dir = new File(appBase, file);
        if (dir.isDirectory()) {
            ContextName cn = new ContextName(file, false);

            if (isServiced(cn.getName()) || deploymentExists(cn.getName()))
                continue;

            results.add(es.submit(new DeployDirectory(this, cn, dir)));
        }
    }

    for (Future<?> result : results) {
        try {
            result.get();
        } catch (Exception e) {
            log.error(sm.getString(
                    "hostConfig.deployDir.threaded.error"), e);
        }
    }
}

1. 通过调试, 能看到:

这里的每一个文件夹, 转换之后, 就是一个 Context .

  

 2. 这里又出现了 Future . 这次提交的是 DeployDirectory, 是 HostConfig 的一个内部类. 看一下他的 run 方法:

private static class DeployDirectory implements Runnable {

    private HostConfig config;
    private ContextName cn;
    private File dir;

    public DeployDirectory(HostConfig config, ContextName cn, File dir) {
        this.config = config;
        this.cn = cn;
        this.dir = dir;
    }

    @Override
    public void run() {
        config.deployDirectory(cn, dir);
    }
}

实际调用的, 还是 HostConfig 的方法:

protected void deployDirectory(ContextName cn, File dir) {
    long startTime = 0;
    // Deploy the application in this directory
    if( log.isInfoEnabled() ) {
        startTime = System.currentTimeMillis();
        log.info(sm.getString("hostConfig.deployDir",
                dir.getAbsolutePath()));
    }

    Context context = null;
    File xml = new File(dir, Constants.ApplicationContextXml);
    File xmlCopy = new File(host.getConfigBaseFile(), cn.getBaseName() + ".xml");

    DeployedApplication deployedApp;
    boolean copyThisXml = isCopyXML();
    boolean deployThisXML = isDeployThisXML(dir, cn);
    try {
        if (deployThisXML && xml.exists()) {
            synchronized (digesterLock) {
                try {
                    //StandardContext
                    context = (Context) digester.parse(xml);
                } catch (Exception e) {
                   ......
                } finally {
                    digester.reset();
                    if (context == null) {
                        context = new FailedContext();
                    }
                }
            }

            if (copyThisXml == false && context instanceof StandardContext) {
                // Host is using default value. Context may override it.
                copyThisXml = ((StandardContext) context).getCopyXML();
            }

            if (copyThisXml) {
                Files.copy(xml.toPath(), xmlCopy.toPath());
                context.setConfigFile(xmlCopy.toURI().toURL());
            } else {
                context.setConfigFile(xml.toURI().toURL());
            }
        } else if (!deployThisXML && xml.exists()) {
            // Block deployment as META-INF/context.xml may contain security
            // configuration necessary for a secure deployment.
            log.error(sm.getString("hostConfig.deployDescriptor.blocked",
                    cn.getPath(), xml, xmlCopy));
            context = new FailedContext();
        } else {
            context = (Context) Class.forName(contextClass).getConstructor().newInstance();
        }

        // 实例化 ContextConfig，作为 LifecycleListener 添加到 Context 容器中，这和 StandardHost 的套路一样，都是使用 XXXConfig
        Class<?> clazz = Class.forName(host.getConfigClass());
        LifecycleListener listener = (LifecycleListener) clazz.getConstructor().newInstance();
        context.addLifecycleListener(listener);

        context.setName(cn.getName());
        context.setPath(cn.getPath());
        context.setWebappVersion(cn.getVersion());
        context.setDocBase(cn.getBaseName());
        // 实例化 StandardContext 之后，为 Host 添加子节点
        // 这里调用的是 StandardHost#addChild()
        host.addChild(context);
    } catch (Throwable t) {
        ......
    } finally {
        ......
    }
    ......
}

这里出现了 host.addChild(context) 方法, 需要回到 StandardHost 类中去看

//org.apache.catalina.core.StandardHost#addChild
public void addChild(Container child) {

    if (!(child instanceof Context))
        throw new IllegalArgumentException
            (sm.getString("standardHost.notContext"));

    //加入一个监听器
    child.addLifecycleListener(new MemoryLeakTrackingListener());

    // Avoid NPE for case where Context is defined in server.xml with only a
    // docBase
    Context context = (Context) child;
    if (context.getPath() == null) {
        ContextName cn = new ContextName(context.getDocBase(), true);
        context.setPath(cn.getPath());
    }
    super.addChild(child);
}

接着看父类中的 addChild()

//org.apache.catalina.core.ContainerBase#addChild
public void addChild(Container child) {
    if (Globals.IS_SECURITY_ENABLED) {
        PrivilegedAction<Void> dp =
            new PrivilegedAddChild(child);
        AccessController.doPrivileged(dp);
    } else {
        addChildInternal(child);
    }
}

private void addChildInternal(Container child) {

    if( log.isDebugEnabled() )
        log.debug("Add child " + child + " " + this);
    synchronized(children) {
        if (children.get(child.getName()) != null)
            throw new IllegalArgumentException("addChild:  Child name '" +
                                               child.getName() +
                                               "' is not unique");
        child.setParent(this);  // May throw IAE
        children.put(child.getName(), child);
    }

    // Start child
    // Don't do this inside sync block - start can be a slow process and
    // locking the children object can cause problems elsewhere
    try {
        if ((getState().isAvailable() ||
                LifecycleState.STARTING_PREP.equals(getState())) &&
                startChildren) {
            //addChild的时候, 调用启动方法, 解析Server.xml时, 不会进此方法
            //StandardHost的时候, 会进此方法, 调用的是 StandardContext.start() -> StandardContext.startInternal()
            //StandardContext的时候, 会进此方法, 调用的是 StandardWrapper.start() -> StandardWrapper.startInternal()
            child.start();
        }
    } catch (LifecycleException e) {
        log.error("ContainerBase.addChild: start: ", e);
        throw new IllegalStateException("ContainerBase.addChild: start: " + e);
    } finally {
        fireContainerEvent(ADD_CHILD_EVENT, child);
    }
}

此时 child 是 StandardContext. 所以会调用

org.apache.catalina.util.LifecycleBase#start() :

public final synchronized void start() throws LifecycleException {

    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;
    }

    if (state.equals(LifecycleState.NEW)) {
        //results.add(startStopExecutor.submit(new StartChild(child)));
        //异步执行 StandardHost.start() 方法时, 会走这里
        init();
    } else if (state.equals(LifecycleState.FAILED)) {
        stop();
    } else if (!state.equals(LifecycleState.INITIALIZED) &&
            !state.equals(LifecycleState.STOPPED)) {
        invalidTransition(Lifecycle.BEFORE_START_EVENT);
    }

    try {
        setStateInternal(LifecycleState.STARTING_PREP, null, false);
        startInternal();
        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 {
            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.
        handleSubClassException(t, "lifecycleBase.startFail", toString());
    }
}

StandardContext 就是在这个方法中进行初始化, 以及启动的.