标签:技术 图片 actor customize ddc post web服务 ext cas
(---------转自网上文章-----------)
我们知道SpringBoot给我们带来了一个全新的开发体验,我们可以把web程序打包成jar包,直接启动,这就得益于SpringBoot内置了容器,可以直接启动。本文将以Tomcat为例,来看看SpringBoot是如何启动Tomcat的,同时也将展开学习Tomcat的源码,了解Tomcat的设计。
用过SpringBoot的人都知道,首先要写一个main方法来启动。
1 @SpringBootApplication 2 public class TomcatDebugApplication { 3 4 public static void main(String[] args) { 5 SpringApplication.run(TomcatDebugApplication.class, args); 6 }
我们直接点击run方法的源码,跟踪下来,发现run方法原来是调用ConfigurableApplicationContext 方法,源码如下:
1 public ConfigurableApplicationContext run(String... args) { 2 StopWatch stopWatch = new StopWatch(); 3 stopWatch.start(); 4 ConfigurableApplicationContext context = null; 5 Collection<SpringBootExceptionReporter> exceptionReporters = new ArrayList<>(); 6 //设置系统属性『java.awt.headless』,为true则启用headless模式支持 7 configureHeadlessProperty(); 8 /* 9 * 通过*SpringFactoriesLoader*检索*META-INF/spring.factories*, 10 * 找到声明的所有SpringApplicationRunListener的实现类并将其实例化, 11 * 之后逐个调用其started()方法,广播SpringBoot要开始执行了 12 */ 13 SpringApplicationRunListeners listeners = getRunListeners(args); 14 //发布应用开始启动事件 15 listeners.starting(); 16 try { 17 //初始化参数 18 ApplicationArguments applicationArguments = new DefaultApplicationArguments( 19 args); 20 /* 21 * 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), 22 * 并遍历调用所有的SpringApplicationRunListener的environmentPrepared()方法,广播Environment准备完毕。 23 */ 24 ConfigurableEnvironment environment = prepareEnvironment(listeners, 25 applicationArguments); 26 configureIgnoreBeanInfo(environment); 27 //打印banner 28 Banner printedBanner = printBanner(environment); 29 //创建应用上下文 30 context = createApplicationContext(); 31 //通过*SpringFactoriesLoader*检索*META-INF/spring.factories*,获取并实例化异常分析器。 32 exceptionReporters = getSpringFactoriesInstances( 33 SpringBootExceptionReporter.class, 34 new Class[] { ConfigurableApplicationContext.class }, context); 35 /* 36 * 为ApplicationContext加载environment,之后逐个执行ApplicationContextInitializer的initialize()方法来进一步封装ApplicationContext, 37 * 并调用所有的SpringApplicationRunListener的contextPrepared()方法,【EventPublishingRunListener只提供了一个空的contextPrepared()方法】, 38 * 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, 39 * 这里就包括通过**@EnableAutoConfiguration**导入的各种自动配置类。 40 */ 41 prepareContext(context, environment, listeners, applicationArguments, 42 printedBanner); 43 //刷新上下文 44 refreshContext(context); 45 //再一次刷新上下文,其实是空方法,可能是为了后续扩展。 46 afterRefresh(context, applicationArguments); 47 stopWatch.stop(); 48 if (this.logStartupInfo) { 49 new StartupInfoLogger(this.mainApplicationClass) 50 .logStarted(getApplicationLog(), stopWatch); 51 } 52 //发布应用已经启动的事件 53 listeners.started(context); 54 /* 55 * 遍历所有注册的ApplicationRunner和CommandLineRunner,并执行其run()方法。 56 * 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。 57 */ 58 callRunners(context, applicationArguments); 59 } 60 catch (Throwable ex) { 61 handleRunFailure(context, ex, exceptionReporters, listeners); 62 throw new IllegalStateException(ex); 63 } 64 65 try { 66 //应用已经启动完成的监听事件 67 listeners.running(context); 68 } 69 catch (Throwable ex) { 70 handleRunFailure(context, ex, exceptionReporters, null); 71 throw new IllegalStateException(ex); 72 } 73 return context; 74 }
其实这个方法可以总结为下面几个步骤:
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
其实上面这段代码,如果只分析Tomcat内容的话,只需要关注两个部分:
上下文是如何创建的,对应方法:createApplicationContext()
上下文是如何刷新的,对应方法:refreshContext(context)
接下来,我们来看看这两个方法做了什么。
1 protected ConfigurableApplicationContext createApplicationContext() { 2 Class<?> contextClass = this.applicationContextClass; 3 if (contextClass == null) { 4 try { 5 switch(this.webApplicationType) { 6 case SERVLET: 7 contextClass = Class.forName("org.springframework.boot.web.servlet.context.AnnotationConfigServletWebServerApplicationContext"); 8 break; 9 case REACTIVE: 10 contextClass = Class.forName("org.springframework.boot.web.reactive.context.AnnotationConfigReactiveWebServerApplicationContext"); 11 break; 12 default: 13 contextClass = Class.forName("org.springframework.context.annotation.AnnotationConfigApplicationContext"); 14 } 15 } catch (ClassNotFoundException var3) { 16 throw new IllegalStateException("Unable create a default ApplicationContext, please specify an ApplicationContextClass", var3); 17 } 18 } 19 20 return (ConfigurableApplicationContext)BeanUtils.instantiateClass(contextClass); 21 }
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AnnotationConfigServletWebServerApplicationContext
类,用图来说明下这个类的关系。
通过这个类图可以知道,这个类继承的是ServletWebServerApplicationContext
,这才是真正的主角,而这个类最终继承了AbstractApplicationContext
,了解完创建上下文的情况后,我们再来看看刷新上下文。
1 //类: SpringApplication 2 3 private void refreshContext(ConfigurableApplicationContext context) { 4 refresh(context); 5 if (this.registerShutdownHook) { 6 try { 7 context.registerShutdownHook(); 8 } 9 catch (AccessControlException ex) { 10 // Not allowed in some environments. 11 } 12 } 13 } 14 15 protected void refresh(ApplicationContext applicationContext) { 16 Assert.isInstanceOf(AbstractApplicationContext.class, applicationContext); 17 ((AbstractApplicationContext) applicationContext).refresh(); 18 }
这里还是直接传递调用本类的refresh(context)
方法,最后强转成父类AbstractApplicationContext
,调用其refresh()
方法,源码如下:
1 //类: AbstractApplicationContext 2 3 public void refresh() throws BeansException, IllegalStateException { 4 synchronized (this.startupShutdownMonitor) { 5 // Prepare this context for refreshing. 6 prepareRefresh(); 7 8 // Tell the subclass to refresh the internal bean factory. 9 ConfigurableListableBeanFactory beanFactory = obtainFreshBeanFactory(); 10 11 // Prepare the bean factory for use in this context. 12 prepareBeanFactory(beanFactory); 13 14 try { 15 // Allows post-processing of the bean factory in context subclasses. 16 postProcessBeanFactory(beanFactory); 17 18 // Invoke factory processors registered as beans in the context. 19 invokeBeanFactoryPostProcessors(beanFactory); 20 21 // Register bean processors that intercept bean creation. 22 registerBeanPostProcessors(beanFactory); 23 24 // Initialize message source for this context. 25 initMessageSource(); 26 27 // Initialize event multicaster for this context. 28 initApplicationEventMulticaster(); 29 30 // Initialize other special beans in specific context subclasses. 31 onRefresh(); 32 33 // Check for listener beans and register them. 34 registerListeners(); 35 36 // Instantiate all remaining (non-lazy-init) singletons. 37 finishBeanFactoryInitialization(beanFactory); 38 39 // Last step: publish corresponding event. 40 finishRefresh(); 41 } 42 43 catch (BeansException ex) { 44 if (logger.isWarnEnabled()) { 45 logger.warn("Exception encountered during context initialization - " + 46 "cancelling refresh attempt: " + ex); 47 } 48 49 // Destroy already created singletons to avoid dangling resources. 50 destroyBeans(); 51 52 // Reset ‘active‘ flag. 53 cancelRefresh(ex); 54 55 // Propagate exception to caller. 56 throw ex; 57 } 58 59 finally { 60 // Reset common introspection caches in Spring‘s core, since we 61 // might not ever need metadata for singleton beans anymore... 62 resetCommonCaches(); 63 } 64 } 65 }
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
1 //类: ServletWebServerApplicationContext 2 3 protected void onRefresh() { 4 super.onRefresh(); 5 try { 6 createWebServer(); 7 } 8 catch (Throwable ex) { 9 throw new ApplicationContextException("Unable to start web server", ex); 10 } 11 } 12 13 private void createWebServer() { 14 WebServer webServer = this.webServer; 15 ServletContext servletContext = getServletContext(); 16 if (webServer == null && servletContext == null) { 17 ServletWebServerFactory factory = getWebServerFactory(); 18 this.webServer = factory.getWebServer(getSelfInitializer()); 19 } 20 else if (servletContext != null) { 21 try { 22 getSelfInitializer().onStartup(servletContext); 23 } 24 catch (ServletException ex) { 25 throw new ApplicationContextException("Cannot initialize servlet context", 26 ex); 27 } 28 } 29 initPropertySources(); 30 }
到这里,终于见到了庐山真面目,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,那就来看看这个工厂的真面目。
根据上图,我们发现工厂类是一个接口,各个具体服务的实现是由各个子类来完成的,所以,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,我们发现其实主要做两件事,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
,这连接器我们勉强能理解(不理解后面会述说),那这个Engine
是什么呢?查看tomcat.getEngine()
的源码:
1 public Engine getEngine() { 2 Service service = getServer().findServices()[0]; 3 if (service.getContainer() != null) { 4 return service.getContainer(); 5 } 6 Engine engine = new StandardEngine(); 7 engine.setName( "Tomcat" ); 8 engine.setDefaultHost(hostname); 9 engine.setRealm(createDefaultRealm()); 10 service.setContainer(engine); 11 return engine; 12 }
根据上面的源码,我们发现,原来这个Engine是容器,继续跟踪源码,找到Container
接口。
上图中,我们看到了4个子接口,分别是Engine、Host、Context、Wrapper
。我们从继承关系上可以知道他们都是容器,那么他们之间的区别是什么呢?那就从它们的注释来分析分析:
1 /** 2 * If used, an Engine is always the top level Container in a Catalina 3 * hierarchy. Therefore, the implementation‘s <code>setParent()</code> method 4 * should throw <code>IllegalArgumentException</code>. 5 * 6 * @author Craig R. McClanahan 7 */ 8 public interface Engine extends Container { 9 //省略代码 10 } 11 12 13 /** 14 * <p> 15 * The parent Container attached to a Host is generally an Engine, but may 16 * be some other implementation, or may be omitted if it is not necessary. 17 * <p> 18 * The child containers attached to a Host are generally implementations 19 * of Context (representing an individual servlet context). 20 * 21 * @author Craig R. McClanahan 22 */ 23 public interface Host extends Container { 24 //省略代码 25 } 26 27 28 /** 29 * <p> 30 * The parent Container attached to a Context is generally a Host, but may 31 * be some other implementation, or may be omitted if it is not necessary. 32 * <p> 33 * The child containers attached to a Context are generally implementations 34 * of Wrapper (representing individual servlet definitions). 35 * <p> 36 * 37 * @author Craig R. McClanahan 38 */ 39 public interface Context extends Container, ContextBind { 40 //省略代码 41 } 42 43 44 /** 45 * <p> 46 * The parent Container attached to a Wrapper will generally be an 47 * implementation of Context, representing the servlet context (and 48 * therefore the web application) within which this servlet executes. 49 * <p> 50 * Child Containers are not allowed on Wrapper implementations, so the 51 * <code>addChild()</code> method should throw an 52 * <code>IllegalArgumentException</code>. 53 * 54 * @author Craig R. McClanahan 55 */ 56 public interface Wrapper extends Container { 57 //省略代码 58 }
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
接着,再看看Tomcat
的源码:
1 //部分代码 2 public class Tomcat { 3 4 //设置连接器 5 public void setConnector(Connector connector) { 6 Service service = getService(); 7 boolean found = false; 8 for (Connector serviceConnector : service.findConnectors()) { 9 if (connector == serviceConnector) { 10 found = true; 11 } 12 } 13 if (!found) { 14 service.addConnector(connector); 15 } 16 } 17 18 //获取service 19 public Service getService() { 20 return getServer().findServices()[0]; 21 } 22 23 //设置Host容器 24 public void setHost(Host host) { 25 Engine engine = getEngine(); 26 boolean found = false; 27 for (Container engineHost : engine.findChildren()) { 28 if (engineHost == host) { 29 found = true; 30 } 31 } 32 if (!found) { 33 engine.addChild(host); 34 } 35 } 36 37 //获取Engine容器 38 public Engine getEngine() { 39 Service service = getServer().findServices()[0]; 40 if (service.getContainer() != null) { 41 return service.getContainer(); 42 } 43 Engine engine = new StandardEngine(); 44 engine.setName( "Tomcat" ); 45 engine.setDefaultHost(hostname); 46 engine.setRealm(createDefaultRealm()); 47 service.setContainer(engine); 48 return engine; 49 } 50 51 //获取server 52 public Server getServer() { 53 54 if (server != null) { 55 return server; 56 } 57 58 System.setProperty("catalina.useNaming", "false"); 59 60 server = new StandardServer(); 61 62 initBaseDir(); 63 64 server.setPort( -1 ); 65 66 Service service = new StandardService(); 67 service.setName("Tomcat"); 68 server.addService(service); 69 return server; 70 } 71 72 //添加context容器 73 public Context addContext(Host host, String contextPath, String contextName, 74 String dir) { 75 silence(host, contextName); 76 Context ctx = createContext(host, contextPath); 77 ctx.setName(contextName); 78 ctx.setPath(contextPath); 79 ctx.setDocBase(dir); 80 ctx.addLifecycleListener(new FixContextListener()); 81 82 if (host == null) { 83 getHost().addChild(ctx); 84 } else { 85 host.addChild(ctx); 86 } 87 return ctx; 88 } 89 90 //添加Wrapper容器 91 public Context addWebapp(Host host, String contextPath, String docBase) { 92 LifecycleListener listener = null; 93 try { 94 Class<?> clazz = Class.forName(getHost().getConfigClass()); 95 listener = (LifecycleListener) clazz.getConstructor().newInstance(); 96 } catch (ReflectiveOperationException e) { 97 // Wrap in IAE since we can‘t easily change the method signature to 98 // to throw the specific checked exceptions 99 throw new IllegalArgumentException(e); 100 } 101 102 return addWebapp(host, contextPath, docBase, listener); 103 } 104 }
阅读Tomcat
的getServer()
方法,我们可以知道,Tomcat
的最顶层就是Server
,也就是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,还可以知道,Engine
容器,一个Service只有一个;根据父子关系,通过setHost()
源码可知,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
)。
根据上面的分析,可以总结出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Ho
st
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. 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步的“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。
标签:技术 图片 actor customize ddc post web服务 ext cas
原文地址:https://www.cnblogs.com/whx7762/p/12304434.html