本文以Tomcat为例,来看看SpringBoot是如何启动Tomcat的,同事展开学习一下Tomcat源码,了解Tomcat设计。
从Main方法说起
用过SpringBoot的都知道,首先要写一个Main方法来启动:
@SpringBootApplication
public class TomcatApplication(String[] args){
SpringApplication.run(TomcatApplication.class, args);
}
我们直接点击run()方法,跟踪后发现run()方法调用的是ConfigurableApplicationContext的run方法,如下:
public ConfigurableApplicationContext run(String... args) {
StopWatch stopWatch = new StopWatch();
stopWatch.start();
ConfigurableApplicationContext context = null;
Collection<SpringBootExceptionReporter> exceptionReporters = new ArrayList();
/**设置系统属性【java.awt.headless】,为true则启动headless模式支持
this.configureHeadlessProperty();通过SpringFactoryLoader检索META- INF/spring.factories,找到声明的所有SpringApplicationRunListener的实现类并将其实例化,而后逐个调用strat()方法,广播SpringBoot要开始执行了
*/
SpringApplicationRunListeners listeners = this.getRunListeners(args);
//发布应用开始启动事件
listeners.starting();
Collection exceptionReporters;
try {
//初始化参数
ApplicationArguments applicationArguments = new DefaultApplicationArguments(args);
/**创建并配置当前SpringBoot应用将要使用的Environment(包括配置要使
用的PropertySource以及Profile),并遍历调用所有的
SpringApplicationRunListener的environmentPrepared()方法,广播
Environment准备完毕。
*/
ConfigurableEnvironment environment = this.prepareEnvironment(listeners, applicationArguments);
this.configureIgnoreBeanInfo(environment);
Banner printedBanner = this.printBanner(environment);
//创建应用上下文
context = this.createApplicationContext();
// 通过SpringFactoriesLoader检索META-INF/spring.factories,获取并实例
化异常分析器
exceptionReporters = this.getSpringFactoriesInstances(SpringBootExceptionReporter.class, new Class[]{ConfigurableApplicationContext.class}, context);
/**为ApplicationContext加载environment,之后逐个执行
ApplicationContextInitializer的initialize()方法来进一步封装
ApplicationContext,并调用所有的SpringApplicationRunListener的contextPrepared()方法,【EventPublishingRunListener只提供了一个空的contextPrepared()方法】,之后初始化IoC容器,并调用SpringApplicationRunListener的contextLoaded()方法,广播ApplicationContext的IoC加载完成,这里就包括通过**@EnableAutoConfiguration**导入的各种自动配置类。*/
this.prepareContext(context, environment, listeners, applicationArguments, printedBanner);
this.refreshContext(context);
this.afterRefresh(context, applicationArguments);
stopWatch.stop();
if (this.logStartupInfo) {
(new StartupInfoLogger(this.mainApplicationClass)).logStarted(this.getApplicationLog(), stopWatch);
}
//发布应用已经启动的事件
listeners.started(context);
this.callRunners(context, applicationArguments);
} catch (Throwable var10) {
/**遍历所有注册的ApplicationRunner和CommandLineRunner,并执行其run()方法。我们可以实现自己的ApplicationRunner或者CommandLineRunner,来对SpringBoot的启动过程进行扩展。
*/
this.handleRunFailure(context, var10, exceptionReporters, listeners);
throw new IllegalStateException(var10);
}
try {
//应用已经启动完成的监听事件
listeners.running(context);
return context;
} catch (Throwable var9) {
this.handleRunFailure(context, var9, exceptionReporters, (SpringApplicationRunListeners)null);
throw new IllegalStateException(var9);
}
}
总结为以下几个步骤:1、配置属性 > 2、获取监听器,发布应用开始启动事件 > 3、初始化输入参数 > 4、配置环境,输出banner > 5、创建上下文 > 6、预处理上下文 > 7、刷新上下文 > 8、再刷新 9、发布应用已经启动事件 > 10、发布应用启动完成事件
如果只分析tomct的话,只需要关注两个内容,上下文是如何创建的,上下文是如何刷新的,分别对应createApplicationContext()和refreshContext(context):
protected ConfigurableApplicationContext createApplicationContext() {
Class<?> contextClass = this.applicationContextClass;
if (contextClass == null) {
try {
switch(this.webApplicationType) {
case SERVLET:
contextClass = Class.forName("org.springframework.boot.web.servlet.context.AnnotationConfigServletWebServerApplicationContext");
break;
case REACTIVE:
contextClass = Class.forName("org.springframework.boot.web.reactive.context.AnnotationConfigReactiveWebServerApplicationContext");
break;
default:
contextClass = Class.forName("org.springframework.context.annotation.AnnotationConfigApplicationContext");
}
} catch (ClassNotFoundException var3) {
throw new IllegalStateException("Unable create a default ApplicationContext, please specify an ApplicationContextClass", var3);
}
}
return (ConfigurableApplicationContext)BeanUtils.instantiateClass(contextClass);
}
这里就是根据WebApplicationType来判断创建哪种类型的Servlet,代码中分别对应着 Web 类型(SERVLET),响应式 Web 类型(REACTIVE),非 Web 类型(default),我们建立的是 Web 类型,所以肯定实例化 DEFAULT_SERVLET_WEB_CONTEXT_CLASS 指定的类,也就是 AnnotationConfigServletWebServerApplicationContext 类,我们来看类图:
通过这个类图我们可以知道,这个类继承的是 ServletWebServerApplicationContext,这就是我们真正的主角,而这个类最终是继承了 AbstractApplicationContext。
接下来看看刷新上下文:
private void refreshContext(ConfigurableApplicationContext context) {
//直接调用刷新方法
this.refresh(context);
if (this.registerShutdownHook) {
try {
context.registerShutdownHook();
} catch (AccessControlException var3) {
}
}
}
protected void refresh(ApplicationContext applicationContext) {
Assert.isInstanceOf(AbstractApplicationContext.class, applicationContext);
((AbstractApplicationContext)applicationContext).refresh();
}
这里还是直接传递调用本类的 refresh(context)方法,最后是强转成父类 AbstractApplicationContext 调用其 refresh()方法,该代码如下:
public void refresh() throws BeansException, IllegalStateException {
synchronized (this.startupShutdownMonitor) {
// Prepare this context for refreshing.
prepareRefresh();
// Tell the subclass to refresh the internal bean factory.
ConfigurableListableBeanFactory beanFactory = obtainFreshBeanFactory();
// Prepare the bean factory for use in this context.
prepareBeanFactory(beanFactory);
try {
// Allows post-processing of the bean factory in context subclasses.
postProcessBeanFactory(beanFactory);
// Invoke factory processors registered as beans in the context.
invokeBeanFactoryPostProcessors(beanFactory);
// Register bean processors that intercept bean creation.
registerBeanPostProcessors(beanFactory);
// Initialize message source for this context.
initMessageSource();
// Initialize event multicaster for this context.
initApplicationEventMulticaster();
// Initialize other special beans in specific context subclasses.这里的意思就是调用各个子类的onRefresh()
onRefresh();
// Check for listener beans and register them.
registerListeners();
// Instantiate all remaining (non-lazy-init) singletons.
finishBeanFactoryInitialization(beanFactory);
// Last step: publish corresponding event.
finishRefresh();
}
catch (BeansException ex) {
if (logger.isWarnEnabled()) {
logger.warn("Exception encountered during context initialization - " +
"cancelling refresh attempt: " + ex);
}
// Destroy already created singletons to avoid dangling resources.
destroyBeans();
// Reset 'active' flag.
cancelRefresh(ex);
// Propagate exception to caller.
throw ex;
}
finally {
// Reset common introspection caches in Spring's core, since we
// might not ever need metadata for singleton beans anymore...
resetCommonCaches();
}
}
}
这里我们看到 onRefresh()方法是调用其子类的实现,根据我们上文的分析,我们这里的子类是 ServletWebServerApplicationContext。
//类:ServletWebServerApplicationContext
protected void onRefresh() {
super.onRefresh();
try {
createWebServer();
}
catch (Throwable ex) {
thrownew ApplicationContextException("Unable to start web server", ex);
}
}
private void createWebServer() {
WebServer webServer = this.webServer;
ServletContext servletContext = getServletContext();
if (webServer == null && servletContext == null) {
ServletWebServerFactory factory = getWebServerFactory();
this.webServer = factory.getWebServer(getSelfInitializer());
}
elseif (servletContext != null) {
try {
getSelfInitializer().onStartup(servletContext);
}
catch (ServletException ex) {
thrownew ApplicationContextException("Cannot initialize servlet context", ex);
}
}
initPropertySources();
}
到这里,其实庐山真面目已经出来了,createWebServer()就是启动 web 服务,但是还没有真正启动 Tomcat,既然 webServer 是通过 ServletWebServerFactory 来获取的,我们就来看看这个工厂的真面目:
走进Tomcat内部
工厂类是一个接口,各个具体服务的实现是由各个子类来实现的,所以我们就去看看 TomcatServletWebServerFactory.getWebServer()的实现。
public WebServer getWebServer(ServletContextInitializer... initializers) {
Tomcat tomcat = new Tomcat();
File baseDir = this.baseDirectory != null ? this.baseDirectory : this.createTempDir("tomcat");
tomcat.setBaseDir(baseDir.getAbsolutePath());
Connector connector = new Connector(this.protocol);
tomcat.getService().addConnector(connector);
this.customizeConnector(connector);
tomcat.setConnector(connector);
tomcat.getHost().setAutoDeploy(false);
this.configureEngine(tomcat.getEngine());
Iterator var5 = this.additionalTomcatConnectors.iterator();
while(var5.hasNext()) {
Connector additionalConnector = (Connector)var5.next();
tomcat.getService().addConnector(additionalConnector);
}
this.prepareContext(tomcat.getHost(), initializers);
return this.getTomcatWebServer(tomcat);
}
根据上面的代码,我们发现其主要做了两件事情,第一件事就是把 Connnctor(我们称之为连接器)对象添加到 Tomcat 中,第二件事就是 configureEngine。
我们查看 tomcat.getEngine()的源码:
public Engine getEngine() {
Service service = getServer().findServices()[0];
if (service.getContainer() != null) {
return service.getContainer();
}
Engine engine = new StandardEngine();
engine.setName( "Tomcat" );
engine.setDefaultHost(hostname);
engine.setRealm(createDefaultRealm());
service.setContainer(engine);
return engine;
}
根据上面的源码,我们发现,这个 Engine 是容器,我们继续跟踪源码,找到 Container 接口,我们看到了 4 个子接口,分别是 Engine,Host,Context,Wrapper,我们从继承关系上看都是容器,他们到底什什么区别呢?
/**
If used, an Engine is always the top level Container in a Catalina
* hierarchy. Therefore, the implementation's <code>setParent()</code> method
* should throw <code>IllegalArgumentException</code>.
*
* @author Craig R. McClanahan
*/
publicinterface Engine extends Container {
//省略代码
}
/**
* <p>
* The parent Container attached to a Host is generally an Engine, but may
* be some other implementation, or may be omitted if it is not necessary.
* </p><p>
* The child containers attached to a Host are generally implementations
* of Context (representing an individual servlet context).
*
* @author Craig R. McClanahan
*/
publicinterface Host extends Container {
//省略代码
}
/*** </p><p>
* The parent Container attached to a Context is generally a Host, but may
* be some other implementation, or may be omitted if it is not necessary.
* </p><p>
* The child containers attached to a Context are generally implementations
* of Wrapper (representing individual servlet definitions).
* </p><p>
*
* @author Craig R. McClanahan
*/
publicinterface Context extends Container, ContextBind {
//省略代码
}
/**</p><p>
* The parent Container attached to a Wrapper will generally be an
* implementation of Context, representing the servlet context (and
* therefore the web application) within which this servlet executes.
* </p><p>
* Child Containers are not allowed on Wrapper implementations, so the
* <code>addChild()</code> method should throw an
* <code>IllegalArgumentException</code>.
*
* @author Craig R. McClanahan
*/
publicinterface Wrapper extends Container {
//省略代码
}
上面的注释翻译过来就是,Engine 是最高级别的容器,其子容器是 Host,Host 的子容器是 Context,Wrapper 是 Context 的子容器,所以这 4 个容器的关系就是父子关系,也就是 Engine>Host>Context>Wrapper。我们再看看 Tomcat 类的源码:
//部分源码,其余部分省略。
publicclass Tomcat {
//设置连接器
public void setConnector(Connector connector) {
Service service = getService();
boolean found = false;
for (Connector serviceConnector : service.findConnectors()) {
if (connector == serviceConnector) {
found = true;
}
}
if (!found) {
service.addConnector(connector);
}
}
//获取service
public Service getService() {
return getServer().findServices()[0];
}
//设置Host容器
public void setHost(Host host) {
Engine engine = getEngine();
boolean found = false;
for (Container engineHost : engine.findChildren()) {
if (engineHost == host) {
found = true;
}
}
if (!found) {
engine.addChild(host);
}
}
//获取Engine容器
public Engine getEngine() {
Service service = getServer().findServices()[0];
if (service.getContainer() != null) {
return service.getContainer();
}
Engine engine = new StandardEngine();
engine.setName( "Tomcat" );
engine.setDefaultHost(hostname);
engine.setRealm(createDefaultRealm());
service.setContainer(engine);
return engine;
}
//获取server
public Server getServer() {
if (server != null) {
return server;
}
System.setProperty("catalina.useNaming", "false");
server = new StandardServer();
initBaseDir();
// Set configuration source
ConfigFileLoader.setSource(new CatalinaBaseConfigurationSource(new File(basedir), null));
server.setPort( -1 );
Service service = new StandardService();
service.setName("Tomcat");
server.addService(service);
return server;
}
//添加Context容器
public Context addContext(Host host, String contextPath, String contextName,
String dir) {
silence(host, contextName);
Context ctx = createContext(host, contextPath);
ctx.setName(contextName);
ctx.setPath(contextPath);
ctx.setDocBase(dir);
ctx.addLifecycleListener(new FixContextListener());
if (host == null) {
getHost().addChild(ctx);
} else {
host.addChild(ctx);
}
//添加Wrapper容器
public static Wrapper addServlet(Context ctx,
String servletName,
Servlet servlet) {
// will do class for name and set init params
Wrapper sw = new ExistingStandardWrapper(servlet);
sw.setName(servletName);
ctx.addChild(sw);
return sw;
}
}
阅读 Tomcat 的 getServer()我们可以知道,Tomcat 的最顶层是 Server,Server 就是 Tomcat 的实例,一个 Tomcat 一个 Server;通过 getEngine()我们可以了解到 Server 下面是 Service,而且是多个,一个 Service 代表我们部署的一个应用,而且我们还可以知道,Engine 容器,一个 service 只有一个;根据父子关系,我们看 setHost()源码可以知道,host 容器有多个;同理,我们发现 addContext()源码下,Context 也是多个;addServlet()表明 Wrapper 容器也是多个,而且这段代码也暗示了,其实 Wrapper 和 Servlet 是一层意思。另外我们根据 setConnector 源码可以知道,连接器(Connector)是设置在 service 下的,而且是可以设置多个连接器(Connector)。
根据上面分析,我们可以小结下:Tomcat 主要包含了 2 个核心组件,连接器(Connector)和容器(Container),用图表示如下:
一个 Tomcat 是一个 Server,一个 Server 下有多个 service,也就是我们部署的多个应用,一个应用下有多个连接器(Connector)和一个容器(Container),容器下有多个子容器,关系用图表示如下:
Engine 下有多个 Host 子容器,Host 下有多个 Context 子容器,Context 下有多个 Wrapper 子容器。
总结
SpringBoot 的启动是通过 new SpringApplication()实例来启动的,启动过程主要做如下几件事情:> 1. 配置属性 > 2. 获取监听器,发布应用开始启动事件 > 3. 初始化输入参数 > 4. 配置环境,输出 banner > 5. 创建上下文 > 6. 预处理上下文 > 7. 刷新上下文 > 8. 再刷新上下文 > 9. 发布应用已经启动事件 > 10. 发布应用启动完成事件。
而启动 Tomcat 就是在第 7 步中“刷新上下文”;Tomcat 的启动主要是初始化 2 个核心组件,连接器(Connector)和容器(Container),一个 Tomcat 实例就是一个 Server,一个 Server 包含多个 Service,也就是多个应用程序,每个 Service 包含多个连接器(Connetor)和一个容器(Container),而容器下又有多个子容器,按照父子关系分别为:Engine,Host,Context,Wrapper,其中除了 Engine 外,其余的容器都是可以有多个。