前言熟悉Tomcat的工程师们,肯定都知道Tomcat是如何启动与停止的。对于startup.sh、startup.bat、shutdown.sh、shutdown.bat等脚本或者批处理命令,大家一定知道改如何使用它,但是它们究竟是如何实现的,尤其是shutdown.sh脚本(或者shutdown.bat)究竟是如何和Tomcat进程通信的呢?本文将通过对Tomcat7.0的源码阅读,深入剖析这一过程。
由于在生产环境中,Tomcat一般部署在linux系统下,所以本文将以startup.sh和shutdown.sh等shell脚本为准,对Tomcat的启动与停止进行分析。
启动过程分析我们启动Tomcat的命令如下:
sh startup.sh
所以,将从shell脚本startup.sh开始分析Tomcat的启动过程。startup.sh的脚本代码见代码清单1。
代码清单1
os400=falsecase"`uname`"inOS400*) os400=true;;esac# resolve links- $0may be a softlinkPRG="$0"while[ -h"$PRG"] ;dols=`ls-ld"$PRG"`
link=`expr"$ls":'.*-> \(.*\)$'`ifexpr"$link":'/.*'> /dev/null;thenPRG="$link"elsePRG=`dirname"$PRG"`/"$link"fidonePRGDIR=`dirname"$PRG"`
EXECUTABLE=catalina.sh# Check that target executable existsif$os400;then#-x will Only work on the os400ifthe files are:
#1. owned by the user
#2. owned by the PRIMARY group of the user
# this will not workifthe user belongsinsecondarygroupsevalelseif[ ! -x"$PRGDIR"/"$EXECUTABLE"];thenecho"Cannot find $PRGDIR/$EXECUTABLE"echo"The file is absent or does not have execute permission"echo"This file is needed to run this program"exit1fifiexec"$PRGDIR"/"$EXECUTABLE"start"$@"
代码清单1中有两个主要的变量,分别是:
PRGDIR:当前shell脚本所在的路径;EXECUTABLE:脚本catalina.sh。根据最后一行代码:exec "$PRGDIR"/"$EXECUTABLE" start "$@",我们知道执行了shell脚本catalina.sh,并且传递参数start。catalina.sh中接收到start参数后的执行的脚本分支见代码清单2。
代码清单2
elif["$1"="start"] ;then# 此处省略参数校验的脚本shifttouch"$CATALINA_OUT"if["$1"="-security"] ;thenif[ $have_tty -eq1];thenecho"Using Security Manager"fishifteval"\"$_RUNjava\"""\"$LOGGING_CONFIG\""$LOGGING_MANAGER $JAVA_OPTS $CATALINA_OPTS \-Djava.endorsed.dirs="\"$JAVA_ENDORSED_DIRS\""-classpath"\"$CLASSPATH\""\-Djava.security.manager \-Djava.security.policy=="\"$CATALINA_BASE/conf/catalina.policy\""\-Dcatalina.base="\"$CATALINA_BASE\""\-Dcatalina.home="\"$CATALINA_HOME\""\-Djava.io.tmpdir="\"$CATALINA_TMPDIR\""\
org.apache.catalina.startup.Bootstrap"$@"start \>>"$CATALINA_OUT"2>&1"&"elseeval"\"$_RUNJAVA\"""\"$LOGGING_CONFIG\""$LOGGING_MANAGER $JAVA_OPTS $CATALINA_OPTS \-Djava.endorsed.dirs="\"$JAVA_ENDORSED_DIRS\""-classpath"\"$CLASSPATH\""\-Dcatalina.base="\"$CATALINA_BASE\""\-Dcatalina.home="\"$CATALINA_HOME\""\-Djava.io.tmpdir="\"$CATALINA_TMPDIR\""\
org.apache.catalina.startup.Bootstrap"$@"start \>>"$CATALINA_OUT"2>&1"&"fiif[ ! -z"$CATALINA_PID"];thenecho$! >"$CATALINA_PID"fiecho"Tomcat started."
从代码清单2可以看出,最终使用java命令执行了org.apache.catalina.startup.Bootstrap类中的main方法,参数也是start。Bootstrap的main方法的实现见代码清单3。
代码清单3
/*** Main method, used for testing only.
*
*@paramargs Command line arguments to be processed*/publicstaticvoidmain(String args[]) {if(daemon ==null) {//Don't set daemon until init() has completedBootstrap bootstrap =newBootstrap();try{
bootstrap.init();
}catch(Throwable t) {
t.printStackTrace();return;
}
daemon=bootstrap;
}try{
String command= "start";if(args.length > 0) {
command= args[args.length - 1];
}if(command.equals("startd")) {
args[args.length- 1] = "start";
daemon.load(args);
daemon.start();
}elseif(command.equals("stopd")) {
args[args.length- 1] = "stop";
daemon.stop();
}elseif(command.equals("start")) {
daemon.setAwait(true);
daemon.load(args);
daemon.start();
}elseif(command.equals("stop")) {
daemon.stopServer(args);
}else{
log.warn("Bootstrap: command \"" + command + "\" does not exist.");
}
}catch(Throwable t) {
t.printStackTrace();
}
}
从代码清单3可以看出,当传递参数start的时候,command等于start,此时main方法的执行步骤如下:
步骤一 初始化Bootstrap
Bootstrap的init方法(见代码清单4)的执行步骤如下:
设置Catalina路径,默认为Tomcat的根目录;初始化Tomcat的类加载器,并设置线程上下文类加载器(具体实现细节,读者可以参考《TOMCAT源码分析——类加载体系》一文);用反射实例化org.apache.catalina.startup.Catalina对象,并且使用反射调用其setParentClassLoader方法,给Catalina对象设置Tomcat类加载体系的顶级加载器(Java自带的三种类加载器除外)。代码清单4
/*** Initialize daemon.*/publicvoidinit()throwsException
{//Set Catalina pathsetCatalinaHome();
setCatalinaBase();
initClassLoaders();
Thread.currentThread().setContextClassLoader(catalinaLoader);
SecurityClassLoad.securityClassLoad(catalinaLoader);//Load our startup class and call its process() methodif(log.isDebugEnabled())
log.debug("Loading startup class");
Class<?> startupClass =catalinaLoader.loadClass
("org.apache.catalina.startup.Catalina");
Object startupInstance=startupClass.newInstance();//Set the shared extensions class loaderif(log.isDebugEnabled())
log.debug("Setting startup class properties");
String methodName= "setParentClassLoader";
Class<?> paramTypes[] =newClass[1];
paramTypes[0] = Class.forName("java.lang.ClassLoader");
Object paramValues[]=newObject[1];
paramValues[0] =sharedLoader;
Method method=startupInstance.getClass().getMethod(methodName, paramTypes);
method.invoke(startupInstance, paramValues);
catalinaDaemon=startupInstance;
}
步骤二 加载、解析server.xml配置文件
当传递参数start的时候,会调用Bootstrap的load方法(见代码清单5),其作用是用反射调用catalinaDaemon(类型是Catalina)的load方法加载和解析server.xml配置文件,具体细节已在《TOMCAT源码分析——SERVER.XML文件的加载与解析》一文中详细介绍,有兴趣的朋友可以选择阅读。
代码清单5
/*** Load daemon.*/privatevoidload(String[] arguments)throwsException {//Call the load() methodString methodName = "load";
Object param[];
Class<?>paramTypes[];if(arguments==null|| arguments.length==0) {
paramTypes=null;
param=null;
}else{
paramTypes=newClass[1];
paramTypes[0] =arguments.getClass();
param=newObject[1];
param[0] =arguments;
}
Method method=catalinaDaemon.getClass().getMethod(methodName, paramTypes);if(log.isDebugEnabled())
log.debug("Calling startup class " +method);
method.invoke(catalinaDaemon, param);
}
步骤三 启动Tomcat
当传递参数start的时候,调用Bootstrap的load方法之后会接着调用start方法(见代码清单6)启动Tomcat,此方法实际是用反射调用了catalinaDaemon(类型是Catalina)的start方法。
代码清单6
/*** Start the Catalina daemon.*/publicvoidstart()throwsException {if( catalinaDaemon==null) init();
Method method= catalinaDaemon.getClass().getMethod("start", (Class [] )null);
method.invoke(catalinaDaemon, (Object [])null);
}
Catalina的start方法(见代码清单7)的执行步骤如下:
验证Server容器是否已经实例化。如果没有实例化Server容器,还会再次调用Catalina的load方法加载和解析server.xml,这也说明Tomcat只允许Server容器通过配置在server.xml的方式生成,用户也可以自己实现Server接口创建自定义的Server容器以取代默认的StandardServer。启动Server容器,有关容器的启动过程的分析可以参考《TOMCAT源码分析——生命周期管理》一文的内容。设置关闭钩子。这么说可能有些不好理解,那就换个说法。Tomcat本身可能由于所在机器断点,程序bug甚至内存溢出导致进程退出,但是Tomcat可能需要在退出的时候做一些清理工作,比如:内存清理、对象销毁等。这些清理动作需要封装在一个Thread的实现中,然后将此Thread对象作为参数传递给Runtime的addShutdownHook方法即可。最后调用Catalina的await方法循环等待接收Tomcat的shutdown命令。如果Tomcat运行正常且没有收到shutdown命令,是不会向下执行stop方法的,当接收到shutdown命令,Catalina的await方法会退出循环等待,然后顺序执行stop方法停止Tomcat。代码清单7
/*** Start a new server instance.*/publicvoidstart() {if(getServer() ==null) {
load();
}if(getServer() ==null) {
log.fatal("Cannot start server. Server instance is not configured.");return;
}longt1 =System.nanoTime();//Start the new servertry{
getServer().start();
}catch(LifecycleException e) {
log.error("Catalina.start: ", e);
}longt2 =System.nanoTime();if(log.isInfoEnabled())
log.info("Server startup in " + ((t2 - t1) / 1000000) + " ms");try{//Register shutdown hookif(useShutdownHook) {if(shutdownHook ==null) {
shutdownHook=newCatalinaShutdownHook();
}
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(logManagerinstanceofClassLoaderLogManager) {
((ClassLoaderLogManager) logManager).setUseShutdownHook(false);
}
}
}catch(Throwable t) {//This will fail on JDK 1.2. Ignoring, as Tomcat can run//fine without the shutdown hook.}if(await) {
await();
stop();
}
}
Catalina的await方法(见代码清单8)实际只是代理执行了Server容器的await方法。
代码清单8
/*** Await and shutdown.*/publicvoidawait() {
getServer().await();
}
以Server的默认实现StandardServer为例,其await方法(见代码清单9)的执行步骤如下:
创建socket连接的服务端对象ServerSocket;循环等待接收客户端发出的命令,如果接收到的命令与SHUTDOWN匹配(由于使用了equals,所以shutdown命令必须是大写的),那么退出循环等待。代码清单9
publicvoidawait() {//Negative values - don't wait on port - tomcat is embedded or we just don't like ports gjaif( port == -2) {//undocumented yet - for embedding apps that are around, alive.return;
}if( port==-1) {while(true) {try{
Thread.sleep(10000);
}catch( InterruptedException ex ) {
}if( stopAwait )return;
}
}//Set up a server socket to wait onServerSocket serverSocket =null;try{
serverSocket=newServerSocket(port, 1,
InetAddress.getByName(address));
}catch(IOException e) {
log.error("StandardServer.await: create[" +address+ ":" +port+ "]: ", e);
System.exit(1);
}//Loop waiting for a connection and a valid commandwhile(true) {//Wait for the next connectionSocket socket =null;
InputStream stream=null;try{
socket=serverSocket.accept();
socket.setSoTimeout(10 * 1000);//Ten secondsstream =socket.getInputStream();
}catch(accessControlException ace) {
log.warn("StandardServer.accept security exception: "
+ace.getMessage(), ace);continue;
}catch(IOException e) {
log.error("StandardServer.await: accept: ", e);
System.exit(1);
}//Read a set of characters from the socketStringBuilder command =newStringBuilder();intexpected = 1024;//Cut off to avoid DoS attackwhile(expected <shutdown.length()) {if(random ==null)
random=newRandom();
expected+= (random.nextInt() % 1024);
}while(expected > 0) {intch = -1;try{
ch=stream.read();
}catch(IOException e) {
log.warn("StandardServer.await: read: ", e);
ch= -1;
}if(ch < 32)//Control character or EOF terminates loop
break;
command.append((char) ch);
expected--;
}//Close the socket now that we are done with ittry{
socket.close();
}catch(IOException e) {//Ignore}//Match against our command stringbooleanmatch =command.toString().equals(shutdown);if(match) {
log.info(sm.getString("standardServer.shutdownViaPort"));
break;
}elselog.warn("StandardServer.await: Invalid command '" +command.toString()+ "' received");
}//Close the server socket and returntry{
serverSocket.close();
}catch(IOException e) {//Ignore}
}
至此,Tomcat启动完毕。很多人可能会问,执行sh shutdown.sh脚本时,是如何与Tomcat进程通信的呢?如果要与Tomcat的ServerSocket通信,socket客户端如何知道服务端的连接地址与端口呢?下面会慢慢说明。
停止过程分析我们停止Tomcat的命令如下:
sh shutdown.sh
所以,将从shell脚本shutdown.sh开始分析Tomcat的停止过程。shutdown.sh的脚本代码见代码清单10。
代码清单10
os400=falsecase"`uname`"in
OS400*) os400=true;;
esac
# resolve links- $0may be a softlink
PRG="$0"while[ -h "$PRG" ] ;dols=`ls -ld "$PRG"`
link=`expr "$ls" : '.*-> \(.*\)$'`ifexpr "$link" : '/.*' > /dev/null; then
PRG="$link"elsePRG=`dirname "$PRG"`/"$link"fi
done
PRGDIR=`dirname "$PRG"`
EXECUTABLE=catalina.sh
# Check that target executable existsif$os400; then
#-x will Only work on the os400ifthe files are:
#1. owned by the user
#2. owned by the PRIMARY group of the user
#thiswill not workifthe user belongs in secondary groups
evalelseif[ ! -x "$PRGDIR"/"$EXECUTABLE"]; then
echo"Cannot find $PRGDIR/$EXECUTABLE"echo"The file is absent or does not have execute permission"echo"This file is needed to run this program"exit1fi
fi
exec"$PRGDIR"/"$EXECUTABLE" stop "$@"
代码清单10和代码清单1非常相似,其中也有两个主要的变量,分别是:
PRGDIR:当前shell脚本所在的路径;EXECUTABLE:脚本catalina.sh。根据最后一行代码:exec "$PRGDIR"/"$EXECUTABLE" stop "$@",我们知道执行了shell脚本catalina.sh,并且传递参数stop。catalina.sh中接收到stop参数后的执行的脚本分支见代码清单11。
代码清单11
elif [ "$1" = "stop"] ; then
#省略参数校验脚本
eval"\"$_RUNJAVA\""$LOGGING_MANAGER $JAVA_OPTS \-Djava.endorsed.dirs="\"$JAVA_ENDORSED_DIRS\"" -classpath "\"$CLASSPATH\""\-Dcatalina.base="\"$CATALINA_BASE\""\-Dcatalina.home="\"$CATALINA_HOME\""\-Djava.io.tmpdir="\"$CATALINA_TMPDIR\""\
org.apache.catalina.startup.Bootstrap"$@" stop
从代码清单11可以看出,最终使用java命令执行了org.apache.catalina.startup.Bootstrap类中的main方法,参数是stop。从代码清单3可以看出,当传递参数stop的时候,command等于stop,此时main方法的执行步骤如下:
步骤一 初始化Bootstrap
已经在启动过程分析中介绍, 不再赘述。
步骤二 停止服务
通过调用Bootstrap的stopServer方法(见代码清单12)停止Tomcat,其实质是用反射调用catalinaDaemon(类型是Catalina)的stopServer方法。
代码清单12
/*** Stop the standalone server.*/publicvoidstopServer(String[] arguments)throwsException {
Object param[];
Class<?>paramTypes[];if(arguments==null|| arguments.length==0) {
paramTypes=null;
param=null;
}else{
paramTypes=newClass[1];
paramTypes[0] =arguments.getClass();
param=newObject[1];
param[0] =arguments;
}
Method method=catalinaDaemon.getClass().getMethod("stopServer", paramTypes);
method.invoke(catalinaDaemon, param);
}
Catalina的stopServer方法(见代码清单13)的执行步骤如下:
创建Digester解析server.xml文件(此处只解析<Server>标签),以构造出Server容器(此时Server容器的子容器没有被实例化);从实例化的Server容器获取Server的socket监听端口和地址,然后创建Socket对象连接启动Tomcat时创建的ServerSocket,最后向ServerSocket发送SHUTDOWN命令。根据代码清单9的内容,ServerSocket循环等待接收到SHUTDOWN命令后,最终调用stop方法停止Tomcat。代码清单13
publicvoidstopServer() {
stopServer(null);
}publicvoidstopServer(String[] arguments) {if(arguments !=null) {
arguments(arguments);
}if( getServer() ==null) {//Create and execute our DigesterDigester digester =createStopDigester();
digester.setClassLoader(Thread.currentThread().getContextClassLoader());
File file=configFile();try{
InputSource is=newInputSource("file://" +file.getAbsolutePath());
FileInputStream fis=newFileInputStream(file);
is.setByteStream(fis);
digester.push(this);
digester.parse(is);
fis.close();
}catch(Exception e) {
log.error("Catalina.stop: ", e);
System.exit(1);
}
}//Stop the existing servertry{if(getServer().getPort()>0) {
Socket socket=newSocket(getServer().getAddress(),
getServer().getPort());
OutputStream stream=socket.getOutputStream();
String shutdown=getServer().getShutdown();for(inti = 0; i < shutdown.length(); i++)
stream.write(shutdown.charAt(i));
stream.flush();
stream.close();
socket.close();
}else{
log.error(sm.getString("catalina.stopServer"));
System.exit(1);
}
}catch(IOException e) {
log.error("Catalina.stop: ", e);
System.exit(1);
}
}
最后,我们看看Catalina的stop方法(见代码清单14)的实现,其执行步骤如下:
将启动过程中添加的关闭钩子移除。Tomcat启动过程辛辛苦苦添加的关闭钩子为什么又要去掉呢?因为关闭钩子是为了在JVM异常退出后,进行资源的回收工作。主动停止Tomcat时调用的stop方法里已经包含了资源回收的内容,所以不再需要这个钩子了。停止Server容器。有关容器的停止内容,请阅读《TOMCAT源码分析——生命周期管理》一文。代码清单14
/*** Stop an existing server instance.*/publicvoidstop() {try{//Remove the ShutdownHook first so that server.stop()//doesn't get invoked twiceif(useShutdownHook) {
Runtime.getRuntime().removeShutdownHook(shutdownHook);//If JULI is being used, re-enable JULI's shutdown to ensure//log messages are not lost jiaanLogManager logManager =LogManager.getLogManager();if(logManagerinstanceofClassLoaderLogManager) {
((ClassLoaderLogManager) logManager).setUseShutdownHook(true);
}
}
}catch(Throwable t) {//This will fail on JDK 1.2. Ignoring, as Tomcat can run//fine without the shutdown hook.}//Shut down the servertry{
getServer().stop();
}catch(LifecycleException e) {
log.error("Catalina.stop", e);
}
}
总结通过对Tomcat源码的分析我们了解到Tomcat的启动和停止都离不开org.apache.catalina.startup.Bootstrap。当停止Tomcat时,已经启动的Tomcat作为socket服务端,停止脚本启动的Bootstrap进程作为socket客户端向服务端发送shutdown命令,两个进程通过共享server.xml里Server标签的端口以及地址信息打通了socket的通信。
如需转载,请标明本文作者及出处——作者:jiaan.gja,本文原创首发:博客园,原文链接:http://www.cnblogs.com/jiaan-geng/p/4872550.html