JavaServer

Servlet Tutorial


This tutorial contains,

  1. An Introduction to Servlets
  2. How to Write a servlet
  3. How to use servletrunner to Run a Servlet

Introduction to Servlets

Servlets are modules that run inside request/response-oriented servers, such as Java-enabled web servers, and extend them in some manner. For example, a servlet might be responsible for taking data in an HTML order-entry form and applying the business logic used to update a company's order database. Servlets are to servers what applets are to browsers.

The Servlet API, which you use to write servlets, assumes nothing about how a servlet is loaded, the server environment in which the servlet runs, or the protocol used to transmit data to and from the user. This allows servlets to be embedded in many different web servers.

Servlets are an effective substitute for CGI scripts: they provide a way to generate dynamic documents that is both easier to write and faster to run. They also address the problem of doing server-side programming with platform-specific APIs. Servlets are developed with the Java Servlet API, a standard Java extension. While it is not part of the core Java framework, which must always be part of all products bearing the Java brand, it will be made available with such products by their vendors as an add-on package. It is already supported by many popular web servers.

Example Uses

A few of the many applications for servlets include,

Servlet Architecture Overview

The central abstraction in the Servlet API is the Servlet interface. All servlets implement this interface, either directly or, more commonly, by extending a class that implements it such as HttpServlet. The Servlet interface provides for methods that manage the servlet and its communications with clients. Servlet writers provide some or all of these methods when developing a servlet.

When a servlet accepts a call from a client, it receives two objects: one is a ServletRequest and the other is a ServletResponse. The ServletRequest class encapsulates the communication from the client to the server, while the ServletResponse class encapsulates the communication from the servlet back to the client.

The ServletRequest interface allows the servlet access to information such as the names of the parameters passed in by the client, the protocol (scheme) being used by the client, and the names of the remote host that made the request and the server that received it. It also provides the servlet with access to the input stream, ServletInputStream, through which the servlet gets data from clients that are using application protocols such as the HTTP POST and PUT methods. Subclasses of ServletRequest allow the servlet to retrieve more protocol-specific data. For example, HttpServletRequest contains methods for accessing HTTP-specific header information.

The ServletResponse interface gives the servlet methods for replying to the client. It allows the servlet to set the content length and mime type of the reply, and provides an output stream, ServletOutputStream, and a Writer through which the servlet can send the reply data. Subclasses of ServletResponse give the servlet more protocol-specific capabilities. For example, HttpServletResponse contains methods that allow the servlet to manipulate HTTP-specific header information.

The classes and interfaces described above make up a basic Servlet. HTTP servlets have some additional objects that provide session-tracking capabilities. The servlet writer can use these APIs to maintain state between the servlet and the client that persists across multiple connections during some time period.

Servlet Lifecycle

Servers load and run servlets, which then accept zero or more requests from clients and return data to them. They can also remove servlets. These are the steps of a servlets lifecycle. The next paragraphs describe each step in more detail, concentrating on concurrency issues.

When a server loads a servlet, it runs the servlet's init method. Even though most servlets are run in multi-threaded servers, there are no concurrency issues during servlet initialization. This is because the server calls the init method once, when it loads the servlet, and will not call it again unless it is reloading the servlet. The server can not reload a servlet until after it has removed the servlet by calling the destroy method. Initialization is allowed to complete before client requests are handled (that is, before the service method is called) or the servlet is destroyed.

After the server loads and initializes the servlet, the servlet is able to handle client requests. It processes them in its service method. Each client's request has its call to the service method run in its own servlet thread: the method receives the client's request, and sends the client its response.

Servlets can run multiple service methods at a time. It is important, therefore, that service methods be written in a thread-safe manner. For example, if a service method might update a field in the servlet object, that access should be synchronized. If, for some reason, a server should not run multiple service methods concurrently, the servlet should implement the SingleThreadModel interface. This interface guarantees that no two threads will execute the servlet's service methods concurrently.

Servlets run until they are removed from the service, for example, at the request of a system administrator. When a server removes a servlet, it runs the servlet's destroy method. The method is run once; the server will not run it again until after it reloads and reinitializes the servlet. When the destroy method runs, however, other threads might be running service requests. If, in cleaning up, it is necessary to access shared resources (such as network connections to be closed), that access should be synchronized.

During a servlet's lifecycle, it is important to write thread-safe code for destroying the servlet and, unless the servlet implements the SingleThreadModel interface, servicing client requests. For more information on writing thread-safe code, please see the Java Tutorial.

Writing the Servlet

Servlets implement the javax.servlet.Servlet interface. While servlet writers can develop servlets by implementing the interface directly, this is usually not required. Because most servlets extend web servers that use the HTTP protocol to interact with clients, the most common way to develop servlets is by specializing the javax.servlet.http.HttpServlet class. This version of the tutorial concentrates on describing this method of writing servlets.

The HttpServlet class implements the Servlet interface by extending the GenericServlet base class, and provides a framework for handling the HTTP protocol. Its service method supports standard HTTP/1.1 requests by dispatching each request to a method designed to handle it.

By default, servlets written by specializing the HttpServlet class can have multiple threads concurrently running its service method. If, for whatever reason, you would like to have only a single thread running a service method at a time, then in addition to extending the HttpServlet, your servlet should also implement the SingleThreadModel interface. This does not involve writing any extra methods, merely declaring that the servlet implements the interface. For example,

public class SurveyServlet extends HttpServlet
                           implements SingleThreadModel {

    /* typical servlet code, with no threading concerns
     * in the service method.  No extra code for the
     * SingleThreadModel interface.  */
    ...

}

Interacting with Clients

Servlet writers who are developing HTTP servlets that specialize the HttpServlet class should override the method or methods designed to handle the HTTP interactions that their servlet will handle. The candidate methods include,

By default, these methods return a BAD_REQUEST (400) error. An example HTTP servlet that handles GET and HEAD requests follows; it specializes the doGet method. A second example is also provided. It handles POST requests from a form by specializing the doPost method.

The HttpServlet's service method, by default, also calls the doOptions method when it receives an OPTIONS request, and doTrace when it receives a TRACE request. The default implementation of doOptions automatically determines what HTTP options are supported and returns that information. The default implementation of doTrace causes a response with a message containing all of the headers sent in the trace request. These methods are not typically overridden.

Whatever method you override, it will take two arguments. The first encapsulates the data from the client, and is an HttpServletRequest. The second encapsulates the response to the client, and is an HttpServletResponse. The following paragraphs discuss their use.

An HttpServletRequest object provides access to HTTP header data, such as any cookies found in the request and the HTTP method with which the request was made. It, of course, allows the you to obtain the arguments that the client sent as part of the request. How you access the client data might depend on the HTTP method of the request.

Note that you should use either the getParameterValues method or one of the methods that allow you to parse the data yourself. They can not be used together in a single request.

For responding to the client, an HttpServletResponse object provides two ways of returning the response data to the user. You can use the writer returned by the getWriter method or the output stream returned by the getOutputStream method. You should use getWriter to return text data to the user, and getOutputStream for binary data.

Before accessing the Writer or OutputStream, HTTP header data should be set. The HttpServletResponse class provides methods to access the header data, such as the content type and encoding, and content length of the response. After you set the headers, you may obtain the writer or output stream and send the body of the response to the user. Closing the writer or output stream after sending the response to the client allows the server to know when the response is complete.

Example of an HTTP Servlet that handles the GET and HEAD methods

/**
 * This is a simple example of an HTTP Servlet.  It responds to the GET
 * and HEAD methods of the HTTP protocol.
 */
public class SimpleServlet extends HttpServlet { 

    public void doGet(HttpServletRequest req, HttpServletResponse res)
        throws ServletException, IOException
    {
        // set header field first
        res.setContentType("text/html");

        // then get the writer and write the response data
        PrintWriter out = res.getWriter();
        out.println("<HEAD><TITLE> SimpleServlet Output</TITLE></HEAD><BODY>");
        out.println("<h1> SimpleServlet Output </h1>");
        out.println("<P>This is output is from SimpleServlet.");
	out.println("</BODY>");
	out.close();
    }

    public String getServletInfo() {
        return "A simple servlet";
    }

}

The example above shows the code for the entire servlet. The doGet method, because it is returning text to the client, uses the HttpServletResponse's getWriter method. It sets the sets the response header field, content type, before writing the body of the response, and closes the writer after writing the response.

In addition to doGet, there is a second method, getServletInfo. More information on the getServletInfo method appears in a later section. Because this servlet is an example shipped with the release, it is already compiled. To try the servlet, run it in the servletrunner.

Example of an HTTP Servlet that handles the POST method

The following example processes data POSTed by a form. (This tutorial assumes that you know HTML. It does not attempt to instruct you on how to develop the forms, only on how to process one using a servlet.) The form looks like this:

<html>
  <head><title>JdcSurvey</title></head>
  <body>
    <form action=http://demo:8080/servlet/survey method=POST>
      <input type=hidden name=survey value=Survey01Results>

      <BR><BR>How Many Employees in your Company?<BR>
        <BR>1-100<input type=radio name=employee value=1-100>
        <BR>100-200<input type=radio name=employee value=100-200>
        <BR>200-300<input type=radio name=employee value=200-300>
        <BR>300-400<input type=radio name=employee value=300-400>
        <BR>500-more<input type=radio name=employee value=500-more>

      <BR><BR>General Comments?<BR>
        <BR><input type=text name=comment>

      <BR><BR>What IDEs do you use?<BR>
        <BR>JavaWorkShop<input type=checkbox name=ide value=JavaWorkShop>
        <BR>J++<input type=checkbox name=ide value=J++>
        <BR>Cafe'<input type=checkbox name=ide value=Cafe'>

      <BR><BR><input type=submit><input type=reset>
    </form>
  </body>
</html>

The servlet writes the form data to a file, and responds to the user with a thank you message. The doPost method of the servlet looks like this:

    /**
     * Write survey results to output file in response to the POSTed
     * form.  Write a "thank you" to the client.     
     */
    public void doPost(HttpServletRequest req, HttpServletResponse res)
	throws ServletException, IOException
    {
        // first, set the "content type" header of the response
	res.setContentType("text/html");

	//Get the response's PrintWriter to return text to the client.
        PrintWriter toClient = res.getWriter();

        try {
            //Open the file for writing the survey results.
            String surveyName = req.getParameterValues("survey")[0];
            FileWriter resultsFile = new FileWriter(resultsDir
	        + System.getProperty("file.separator")
	        + surveyName + ".txt", true);
            PrintWriter toFile = new PrintWriter(resultsFile);

	    // Get client's form data & store it in the file
            toFile.println("<BEGIN>");
            Enumeration values = req.getParameterNames();
            while(values.hasMoreElements()) {
                String name = (String)values.nextElement();
		String value = req.getParameterValues(name)[0];
                if(name.compareTo("submit") != 0) {
                    toFile.println(name + ": " + value);
                }
            }
            toFile.println("<END>");

	    //Close the file.
            resultsFile.close();

	    // Respond to client with a thank you
	    toClient.println("<html>");
	    toClient.println("<title>Thank you!</title>");
            toClient.println("Thank you for participating");
	    toClient.println("</html>");

        } catch(IOException e) {
            e.printStackTrace();
            toClient.println(
		"A problem occured while recording your answers.  "
		+ "Please try again.");
        }

        // Close the writer; the response is done.
	toClient.close();
    }

The doPost method uses the getParameterNames and getParameterValues methods to get the form data. Because it returns text to the client, doPost calls the getWriter method. It sets the sets the response header field, content type, before writing the body of the response, and closes the writer when the response is complete.

Lifecycle Methods

Servlets that manage resources do so by overriding the lifecycle methods init and destroy. These servlets might need to be given arguments at startup, in order to initialize correctly.

Overriding the Init Method

During initialization, the servlet should prepare the resources it manages, to ready the servlet for accepting service requests. It can do this without regard for multi-threading concerns, since there is only a single thread running on behalf of the servlet during initialization. As soon as the init method returns, the servlet can receive client requests. If, for some reason, the servlet's required resources can not be made available (for example, a required network connection can not be established), or some other initialization error occurs that would make it impossible for the servlet to handle requests, the init method should throw an UnavailableException exception.

The init method takes a ServletConfig object as a parameter. The method should save this object, so that it can be returned by the getServletConfig method. The simplest way to do this is to have the new init method call super.init. If you do not do this, you should store the ServletConfig object yourself, and override the getServletConfig method so that it can obtain the object from its new location.

An example init method follows. It is the init method from the Survey Servlet, which accepts input from a form and stores it in a file. In order store the survey information, it needs a directory. It receives the directory as an initialization parameter; initialization parameters are discussed in the next section.

    public void init(ServletConfig config)
	throws ServletException
    {
	super.init(config);

        //Store the directory that will hold the survey-results files
        resultsDir = getInitParameter("resultsDir");

        //If no directory was provided, can't handle clients
	if (resultsDir == null) {
	    throw new UnavailableException (this,
		"Not given a directory to write survey results!");
	}
    }

As you can see, this init method calls the super.init method to manage the ServletConfig object. The init method also sets a field, resultsDir, with the directory name that is provided as an initialization parameter. If no directory name was provided, the servlet throws an unavailable exception. If the init method completes successfully, the servlet can then handle client requests.

Initialization Parameters

The specification of initialization parameters is server-specific. For example, they are specified with a property when a servlet is run with the servlet runner. This tutorial contains a general explanation of properties, and how to create them.

However the initialization parameters are specified, they are always obtained the same way: with the getInitParameter method. This method takes the parameter name as an argument. The example init method calls getInitParameter. If, for some reason, you need to get the parameter names, you can get them with the getParameterNames method.

Overriding the Destroy Method

When a server unloads a servlet, it calls the servlet's destroy method. The destroy method should undo any initialization work and synchronize persistent state with the current in-memory state. This section begins with a description of how to write a simple destroy method, then describes how to structure a servlet if threads running its service method might still be running when the destroy method is called.

Though it is often the case that a servlet that overrides the init method must also override the destroy method to undo that initialization, this is not required. For example, the phone servlet, whose init method is used as an example, does not have a corresponding destroy method. Because initialization involves reading a file and using its contents to initialize a shared data structure, there is no work to undo when the server is finished with the servlet.

For many servlets, however, there is initialization work that must be undone. For example, assume there is a servlet that opens a database connection during initialization. Its destroy method, shown as an example below, would close that connection.

    /**
     * Cleans up database connection
     */
    public void destroy() {
        try {
            con.close();
        } catch (SQLException e) {
            while(e != null) {
		log("SQLException: " + e.getSQLState() + '\t' +
		    e.getMessage() + '\t' +
		    e.getErrorCode() + '\t');
		e = e.getNextException();
            }
        } catch (Exception e) {
	    e.printStackTrace();
	}
    }
Coping with Service Threads at Servlet Termination

When a server removes a servlet, it typically calls destroy after all service calls have been completed, or a server-specific number of seconds have passed, whichever comes first. If your servlet has operations that take a long time to run (that is, they may run longer than the server's grace period), then threads could still be running when destroy is called. The servlet writer is responsible for making sure that any threads still handling client requests complete; the remainder of this section describes a technique for doing this.

A servlet with potentially long-running service requests should keep track of how many service methods are currently running. Its long-running methods should periodically poll to make sure that they should continue to run. If the servlet is being destroyed, then the long-running method should stop working, clean up if necessary, and return.

For example, the instance variable that counts the number of service methods running could be called serviceCounter, and the indicator of whether the servlet is being destroyed could be an instance variable called shuttingDown. Each variable should have its own set of access methods:

public ShutdownExample extends HttpServlet {
    private int serviceCounter = 0;
    private Boolean shuttingDown;
    ...
    //Access methods for serviceCounter
    protected synchronized void enteringServiceMethod() {
	serviceCounter++;
    }
    protected synchronized void leavingServiceMethod() {
        serviceCounter--;
    }
    protected synchronized int numServices() {
	return serviceCounter;
    }
    //Access methods for shuttingDown
    protected setShuttingDown(Boolean flag) {
	shuttingDown = flag;
    }
    protected Boolean isShuttingDown() {
	return shuttingDown;
    }
}

The service method should increment the service counter each time it is entered and decrement it each time it returns:

    protected void service(HttpServletRequest req, HttpServletResponse resp)
        throws ServletException, IOException
    {
	enteringServiceMethod();
	try {
            super.service(req, resp);
        } finally {
            leavingServiceMethod();
        }
    }

The destroy method should check the serviceCounter, and if there are any long-running methods, set the shuttingDown variable. This variable will let the threads still handling client requests know that it is time to shut down. The destroy method should then wait for the service methods to complete, in order to provide a clean shutdown.

    public void destroy() {
        /* Check to see whether there are still service methods running,
	 * and if there are, tell them to stop. */
	if (numServices() > 0) {
	    setShuttingDown(true);
        }

	/* Wait for the service methods to stop.  */
	while(numService() > 0) {
            try {
                thisThread.sleep(interval);
            } catch (InterruptedException e) {
            }
        }
    }
The long-running methods should check this variable, and interrupt their work if neceesary:
    public void doPost(...) {
        ...
	for(i = 0; ((i < lotsOfStuffToDo) && !isShuttingDown()); i++) {
	    try {
		partOfLongRunningOperation(i);
	    } catch (InterruptedException e) {
            }
        }
    }

Providing Information about the Servlet

Some applets and applications, for example, the Java Web Server Administration Tool, display information about a servlet. This information can include a short description of the purpose of the servlet, its author, and perhaps its version number. The Servlet API provides a method to return this information, getServletInfo. By default, this method returns null. While servlet writers are not required to override this method, it is strongly recommended. The simple servlet, shown as an example earlier, overrides this method:

/**
 * This is a simple example of an HTTP Servlet.  It responds to the GET
 * and HEAD methods of the HTTP protocol.
 */
public class SimpleServlet extends HttpServlet { 

   ...

    public String getServletInfo() {
        return "A simple servlet";
    }
}

How to use servletrunner to Run a Servlet

Once you have written your servlet, you can run it in many web servers, or in the servletrunner. Where ever you decide to run your servlet, there are certain pieces of data that you might want or need to specify. When you are using servletrunner you do this with properties. The next section describes a servlet's properties, and how to store them. Following that, there is a section on how to run servlets in servletrunner.

Properties

Properties are key-value pairs, used for the configuration, creation, and initialization of a servlet. For example, servlet.phone.code=PhoneServlet is a property whose key is servlet.phone.code and whose value is PhoneServlet.

There are two properties for servlets. One is servlet.name.code, whose value is the servlet's class name. The other property is servlet.name.initargs, whose value holds the initialization parameters for the servlet. Both properties, servlet.name.code and servlet.name.initargs, are discussed in more detail below.

Using the code Property

The servlet.name.code property names your servlet by associating its name with its class. If your servlet uses initialization parameters, this property is required. It allows the server to associate the servlet object with its initialization arguments: they both have the same name. Even if your servlet does not use initialization parameters, it is recommended that it have this property, so that clients can access the servlet using its name.

Syntax of the Initargs Property

The value of the servlet.name.initArgs property holds the servlet's initialization parameters. The syntax of a single parameter is parameterName=parameterValue. The entire property (the entire key-value pair) must be a single logical line. For readability, you can use the backquote syntax to allow the property to span multiple lines in the file. For example, the argument to the phone servlet looks like this:

servlet.phone.initArgs=\
        phonelist=servlets/phonelist

If there are multiple initialization parameters, they are specified as a comma-delimited list. For example, the arguments to a database demo could look like this:

servlet.dbdemo.initArgs=\
	username=fill_in_the_user,\
	password=fill_in_the_password,\
	owner=fill_in_the_name
The Property File

Properties are stored in a file that is, by default, called "servlet.properties", though you can specify another name when servletrunner is started. The file should hold the properties for all the servlets that servletrunner will run. It should be plain text; you can create it in an editor. Here is an excerpt of a servlet.properties file as an example:

# phone servlet (sample.html)
servlet.phone.code=PhoneServlet
servlet.phone.initArgs=\
      phonelist=servlets/phonelist

# bulletin board servlet
servlet.bboard.code=BBoardServlet

# order entry servlet
servlet.dbdemo.code=OrderEntryServlet
servlet.dbdemo.initArgs=\
	username=fill_in_the_user,\
	password=fill_in_the_password,\
	owner=fill_in_the_name

Using Servlet Runner

If you would like to run your servlet in a web server, please see that server's documentation for instructions. This section explains how to run the servlet in the servletrunner utility that comes with this release.

The servletrunner is a small utility, intended for testing. It is multithreaded, so it can run more than one servlet. It can be used, therefore, to run multiple servlets simultaneously, or to test one servlet that calls other servlets in order to satisfy client requests. Unlike some web servers, it does not automatically reload servlets when they are updated. Because it is small, however, there is very little overhead associated with stopping and restarting it in order to use a new version of a servlet.

The servletrunner is in the <JDK>/bin directory. Invoking it with the -help flag shows a usage message without running it:

% ./bin/servletrunner -help
Usage: servletrunner [options]
Options:
  -p port     the port number to listen on
  -b backlog  the listen backlog
  -m max      maximum number of connection handlers
  -t timeout  connection timeout in milliseconds
  -d dir      servlet directory
  -r root     document root directory
  -s filename servlet property file name
  -v          verbose output
%

In order to see the default values of these options, you can call servletrunner with the -v switch. This will, however, start the servlet runner. Just stop it after you have obtained the information, if you are not ready to run it yet, or want it to run with something other than the default values.

% ./bin/servletrunner -v
Server settings:
  port = 8080
  backlog = 50
  max handlers = 100
  timeout = 5000
  servlet dir = .
  document dir = .
  servlet propfile = .:servlet.properties
 
 

Once the servletrunner is executing, you run servlets by calling them directly in your browser, or as the forms example shows, by using a form that calls a servlet to process its data. The URL for a servlet has the following general form:

http://machine-name:port/servlet/servlet-name
where servlet-name corresponds to the name you have given your servlet. For example, to run the Phone Servlet, which has the property servlet.phone.code=PhoneServlet, you would use the following URL. (It assumes that servletrunner is running on your machine, localhost, at port 8080, and that the phone servlet is located in the servlet directory provided to servletrunner at startup:
http://localhost:8080/servlet/phone
Another example, the survey servlet, is run as a result of submitting a form. The URL used for the servlet by the survey-form in the HTML file is:
http://demo:8080/servlet/survey
Here is the output of the survey servlet:

output of the survey servlet


Top
java-server-feedback@java.sun.com