Thanks for theory but this is supposed to be my first servlet so I need
concrete answer if I have this servlet:

// import the JavaMail packages
import javax.mail.*;
import javax.mail.internet.*;

// import the servlet packages
import javax.servlet.*;
import javax.servlet.http.*;

// import misc classes that we need
import java.util.*;
import java.io.*;

public class SendMailServlet extends HttpServlet {
   String smtpServer;

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

       // get the SMTP server from the servlet properties
       smtpServer = config.getInitParameter("smtpServer");
   }

   public void doPost(HttpServletRequest req, HttpServletResponse res)
                      throws ServletException, IOException
   {
       // get the message parameters from the HTML page
       String from = req.getParameter("from");
       String to = req.getParameter("to");
       String subject = req.getParameter("subject");
       String text = req.getParameter("text");

       PrintWriter out = res.getWriter();
       res.setContentType("text/html");

       try {
           // set the SMTP host property value
           Properties properties = System.getProperties();
           properties.put("mail.smtp.host", smtpServer);

           // create a JavaMail session
           Session session = Session.getInstance(properties, null);

           // create a new MIME message
           MimeMessage message = new MimeMessage(session);

           // set the from address
           Address fromAddress = new InternetAddress(from);
           message.setFrom(fromAddress);

           // set the to address
           if (to != null) {
               Address[] toAddress = InternetAddress.parse(to);
               message.setRecipients(Message.RecipientType.TO, toAddress);
           }
           else
               throw new MessagingException("No \"To\" address specified");

           // set the subject
           message.setSubject(subject);

           // set the message body
           message.setText(text);

           // send the message
           Transport.send(message);

           out.println("Message sent successfully.");
       }
       catch (AddressException e) {
           out.println("Invalid e-mail address.<br>" + e.getMessage());
       }
       catch (SendFailedException e) {
           out.println("Send failed.<br>" + e.getMessage());
       }
       catch (MessagingException e) {
           out.println("Unexpected error.<br>" + e.getMessage());
       }
   }
}

And this tomcat -> server.xml:

<?xml version="1.0" encoding="UTF-8"?>
<!-- TOMCAT ADMIN login je ide a lozinka se nalazi u
C:\Documents and Settings\Dalibor
®upan\.netbeans\5.0\jakarta-tomcat-5.5.9_base\conf
-->
<!-- Example Server Configuration File -->
<!-- Note that component elements are nested corresponding to their
    parent-child relationships with each other -->
<!-- A "Server" is a singleton element that represents the entire JVM,
    which may contain one or more "Service" instances.  The Server
    listens for a shutdown command on the indicated port.

    Note:  A "Server" is not itself a "Container", so you may not
    define subcomponents such as "Valves" or "Loggers" at this level.
-->
<Server port="8025" shutdown="SHUTDOWN">
 <!-- Comment these entries out to disable JMX MBeans support used for the
      administration web application -->
 <Listener className="org.apache.catalina.mbeans.ServerLifecycleListener"/>
 <Listener
className="org.apache.catalina.mbeans.GlobalResourcesLifecycleListener"/>
 <Listener
className="org.apache.catalina.storeconfig.StoreConfigLifecycleListener"/>
 <!-- Global JNDI resources -->
 <GlobalNamingResources>
   <!-- Test entry for demonstration purposes -->
   <Environment name="simpleValue" type="java.lang.Integer" value="30"/>
   <!-- Editable user database that can also be used by
        UserDatabaseRealm to authenticate users -->
   <Resource auth="Container" description="User database that can be
updated and saved"
factory="org.apache.catalina.users.MemoryUserDatabaseFactory"
name="UserDatabase" pathname="conf/tomcat-users.xml"
type="org.apache.catalina.UserDatabase"/>
 </GlobalNamingResources>
 <!-- A "Service" is a collection of one or more "Connectors" that share
      a single "Container" (and therefore the web applications visible
      within that Container).  Normally, that Container is an "Engine",
      but this is not required.

      Note:  A "Service" is not itself a "Container", so you may not
      define subcomponents such as "Valves" or "Loggers" at this level.
  -->
 <!-- Define the Tomcat Stand-Alone Service -->
 <Service name="Catalina">
   <!-- A "Connector" represents an endpoint by which requests are received
        and responses are returned.  Each Connector passes requests on to
the
        associated "Container" (normally an Engine) for processing.

        By default, a non-SSL HTTP/1.1 Connector is established on port
8080.
        You can also enable an SSL HTTP/1.1 Connector on port 8443 by
        following the instructions below and uncommenting the second
Connector
        entry.  SSL support requires the following steps (see the SSL
Config
        HOWTO in the Tomcat 5 documentation bundle for more detailed
        instructions):
        * If your JDK version 1.3 or prior, download and install JSSE 1.0.2
or
          later, and put the JAR files into "$JAVA_HOME/jre/lib/ext".
        * Execute:
            %JAVA_HOME%\bin\keytool -genkey -alias tomcat -keyalg RSA
(Windows)
            $JAVA_HOME/bin/keytool -genkey -alias tomcat -keyalg RSA
(Unix)
          with a password value of "changeit" for both the certificate and
          the keystore itself.

        By default, DNS lookups are enabled when a web application calls
        request.getRemoteHost().  This can have an adverse impact on
        performance, so you can disable it by setting the
        "enableLookups" attribute to "false".  When DNS lookups are
disabled,
        request.getRemoteHost() will return the String version of the
        IP address of the remote client.
   -->
   <!-- Define a non-SSL HTTP/1.1 Connector on port 8080 -->
   <Connector URIEncoding="utf-8" acceptCount="100"
connectionTimeout="20000" disableUploadTimeout="true" enableLookups="false"
maxHttpHeaderSize="8192" maxSpareThreads="75" maxThreads="150"
minSpareThreads="25" port="8084" redirectPort="8443"/>
   <!-- Note : To disable connection timeouts, set connectionTimeout value
    to 0 -->
   <!-- Note : To use gzip compression you could set the following
properties :

     compression="on"
     compressionMinSize="2048"
     noCompressionUserAgents="gozilla, traviata"
     compressableMimeType="text/html,text/xml"
-->
   <!-- Define a SSL HTTP/1.1 Connector on port 8443 -->
   <!--
   <Connector port="8443" maxHttpHeaderSize="8192"
              maxThreads="150" minSpareThreads="25" maxSpareThreads="75"
              enableLookups="false" disableUploadTimeout="true"
              acceptCount="100" scheme="https" secure="true"
              clientAuth="false" sslProtocol="TLS" />
   -->
   <!-- Define an AJP 1.3 Connector on port 8009 -->
   <Connector enableLookups="false" port="8009" protocol="AJP/1.3"
redirectPort="8443"/>
   <!-- Define a Proxied HTTP/1.1 Connector on port 8082 -->
   <!-- See proxy documentation for more information about using this. -->
   <!--
   <Connector port="8082"
              maxThreads="150" minSpareThreads="25" maxSpareThreads="75"
              enableLookups="false" acceptCount="100"
connectionTimeout="20000"
              proxyPort="80" disableUploadTimeout="true" />
   -->
   <!-- An Engine represents the entry point (within Catalina) that
processes
        every request.  The Engine implementation for Tomcat stand alone
        analyzes the HTTP headers included with the request, and passes
them
        on to the appropriate Host (virtual host). -->
   <!-- You should set jvmRoute to support load-balancing via AJP ie :
   <Engine name="Standalone" defaultHost="localhost" jvmRoute="jvm1">
   -->
   <!-- Define the top level container in our container hierarchy -->
   <Engine defaultHost="localhost" name="Catalina">
     <!-- The request dumper valve dumps useful debugging information about
          the request headers and cookies that were received, and the
response
          headers and cookies that were sent, for all requests received by
          this instance of Tomcat.  If you care only about requests to a
          particular virtual host, or a particular application, nest this
          element inside the corresponding <Host> or <Context> entry
instead.

          For a similar mechanism that is portable to all Servlet 2.4
          containers, check out the "RequestDumperFilter" Filter in the
          example application (the source for this filter may be found in
          "$CATALINA_HOME/webapps/examples/WEB-INF/classes/filters").

          Request dumping is disabled by default.  Uncomment the following
          element to enable it. -->
     <!--
     <Valve className="org.apache.catalina.valves.RequestDumperValve"/>
     -->
     <!-- Because this Realm is here, an instance will be shared
globally -->
     <!-- This Realm uses the UserDatabase configured in the global JNDI
          resources under the key "UserDatabase".  Any edits
          that are performed against this UserDatabase are immediately
          available for use by the Realm.  -->
     <Realm className="org.apache.catalina.realm.UserDatabaseRealm"
resourceName="UserDatabase"/>
     <!-- Comment out the old realm but leave here for now in case we
          need to go back quickly -->
     <!--
     <Realm className="org.apache.catalina.realm.MemoryRealm" />
     -->
     <!-- Replace the above Realm with one of the following to get a Realm
          stored in a database and accessed via JDBC -->
     <!--
     <Realm  className="org.apache.catalina.realm.JDBCRealm"
            driverName="org.gjt.mm.mysql.Driver"
         connectionURL="jdbc:mysql://localhost/authority"
        connectionName="test" connectionPassword="test"
             userTable="users" userNameCol="user_name"
userCredCol="user_pass"
         userRoleTable="user_roles" roleNameCol="role_name" />
     -->
     <!--
     <Realm  className="org.apache.catalina.realm.JDBCRealm"
            driverName="oracle.jdbc.driver.OracleDriver"
         connectionURL="jdbc:oracle:thin:@ntserver:1521:ORCL"
        connectionName="scott" connectionPassword="tiger"
             userTable="users" userNameCol="user_name"
userCredCol="user_pass"
         userRoleTable="user_roles" roleNameCol="role_name" />
     -->
     <!--
     <Realm  className="org.apache.catalina.realm.JDBCRealm"
            driverName="sun.jdbc.odbc.JdbcOdbcDriver"
         connectionURL="jdbc:odbc:CATALINA"
             userTable="users" userNameCol="user_name"
userCredCol="user_pass"
         userRoleTable="user_roles" roleNameCol="role_name" />
     -->
     <!-- Define the default virtual host
          Note: XML Schema validation will not work with Xerces 2.2.
      -->
     <Host appBase="webapps" autoDeploy="false" name="localhost"
unpackWARs="true" xmlNamespaceAware="false" xmlValidation="false">
       <!-- Defines a cluster for this node,
            By defining this element, means that every manager will be
changed.
            So when running a cluster, only make sure that you have webapps
in there
            that need to be clustered and remove the other ones.
            A cluster has the following parameters:

            className = the fully qualified name of the cluster class

            name = a descriptive name for your cluster, can be anything

            mcastAddr = the multicast address, has to be the same for all
the nodes

            mcastPort = the multicast port, has to be the same for all the
nodes

            mcastBindAddr = bind the multicast socket to a specific address

            mcastTTL = the multicast TTL if you want to limit your
broadcast

            mcastSoTimeout = the multicast readtimeout

            mcastFrequency = the number of milliseconds in between sending
a "I'm alive" heartbeat

            mcastDropTime = the number a milliseconds before a node is
considered "dead" if no heartbeat is received

            tcpThreadCount = the number of threads to handle incoming
replication requests, optimal would be the same amount of threads as nodes

            tcpListenAddress = the listen address (bind address) for TCP
cluster request on this host,
                               in case of multiple ethernet cards.
                               auto means that address becomes
                               InetAddress.getLocalHost().getHostAddress()

            tcpListenPort = the tcp listen port

            tcpSelectorTimeout = the timeout (ms) for the Selector.select()
method in case the OS
                                 has a wakup bug in java.nio. Set to 0 for
no timeout

            printToScreen = true means that managers will also print to
std.out

            expireSessionsOnShutdown = true means that

            useDirtyFlag = true means that we only replicate a session
after setAttribute,removeAttribute has been called.
                           false means to replicate the session after each
request.
                           false means that replication would work for the
following piece of code: (only for SimpleTcpReplicationManager)
                           <%
                           HashMap map =
(HashMap)session.getAttribute("map");
                           map.put("key","value");
                           %>
            replicationMode = can be either 'pooled', 'synchronous' or
'asynchronous'.
                              * Pooled means that the replication happens
using several sockets in a synchronous way. Ie, the data gets replicated,
then the request return. This is the same as the 'synchronous' setting
except it uses a pool of sockets, hence it is multithreaded. This is the
fastest and safest configuration. To use this, also increase the nr of tcp
threads that you have dealing with replication.
                              * Synchronous means that the thread that
executes the request, is also the
                              thread the replicates the data to the other
nodes, and will not return until all
                              nodes have received the information.
                              * Asynchronous means that there is a specific
'sender' thread for each cluster node,
                              so the request thread will queue the
replication request into a "smart" queue,
                              and then return to the client.
                              The "smart" queue is a queue where when a
session is added to the queue, and the same session
                              already exists in the queue from a previous
request, that session will be replaced
                              in the queue instead of replicating two
requests. This almost never happens, unless there is a
                              large network delay.
       -->
       <!--
           When configuring for clustering, you also add in a valve to
catch all the requests
           coming in, at the end of the request, the session may or may not
be replicated.
           A session is replicated if and only if all the conditions are
met:
           1. useDirtyFlag is true or setAttribute or removeAttribute has
been called AND
           2. a session exists (has been created)
           3. the request is not trapped by the "filter" attribute

           The filter attribute is to filter out requests that could not
modify the session,
           hence we don't replicate the session after the end of this
request.
           The filter is negative, ie, anything you put in the filter, you
mean to filter out,
           ie, no replication will be done on requests that match one of
the filters.
           The filter attribute is delimited by ;, so you can't escape out
; even if you wanted to.

           filter=".*\.gif;.*\.js;" means that we will not replicate the
session after requests with the URI
           ending with .gif and .js are intercepted.

           The deployer element can be used to deploy apps cluster wide.
           Currently the deployment only deploys/undeploys to working
members in the cluster
           so no WARs are copied upons startup of a broken node.
           The deployer watches a directory (watchDir) for WAR files when
watchEnabled="true"
           When a new war file is added the war gets deployed to the local
instance,
           and then deployed to the other instances in the cluster.
           When a war file is deleted from the watchDir the war is
undeployed locally
           and cluster wide
       -->
       <!--
       <Cluster
className="org.apache.catalina.cluster.tcp.SimpleTcpCluster"
                managerClassName="org.apache.catalina.cluster.session.DeltaManager"
                expireSessionsOnShutdown="false"
                useDirtyFlag="true"
                notifyListenersOnReplication="true">

           <Membership
               className="org.apache.catalina.cluster.mcast.McastService"
               mcastAddr="228.0.0.4"
               mcastPort="45564"
               mcastFrequency="500"
               mcastDropTime="3000"/>

           <Receiver
               className="org.apache.catalina.cluster.tcp.ReplicationListener"
               tcpListenAddress="auto"
               tcpListenPort="4001"
               tcpSelectorTimeout="100"
               tcpThreadCount="6"/>

           <Sender
               className="org.apache.catalina.cluster.tcp.ReplicationTransmitter"
               replicationMode="pooled"
               ackTimeout="15000"/>

           <Valve
className="org.apache.catalina.cluster.tcp.ReplicationValve"
                  filter=".*\.gif;.*\.js;.*\.jpg;.*\.png;.*\.htm;.*\.html;.*\.css;.*\.txt;"/>

           <Deployer
className="org.apache.catalina.cluster.deploy.FarmWarDeployer"
                     tempDir="/tmp/war-temp/"
                     deployDir="/tmp/war-deploy/"
                     watchDir="/tmp/war-listen/"
                     watchEnabled="false"/>
       </Cluster>
       -->
       <!-- Normally, users must authenticate themselves to each web app
            individually.  Uncomment the following entry if you would like
            a user to be authenticated the first time they encounter a
            resource protected by a security constraint, and then have that
            user identity maintained across *all* web applications
contained
            in this virtual host. -->
       <!--
       <Valve className="org.apache.catalina.authenticator.SingleSignOn" />
       -->
       <!-- Access log processes all requests for this virtual host.  By
            default, log files are created in the "logs" directory relative
to
            $CATALINA_HOME.  If you wish, you can specify a different
            directory with the "directory" attribute.  Specify either a
relative
            (to $CATALINA_HOME) or absolute path to the desired directory.
       -->
       <!--
       <Valve className="org.apache.catalina.valves.AccessLogValve"
                directory="logs"  prefix="localhost_access_log."
suffix=".txt"
                pattern="common" resolveHosts="false"/>
       -->
       <!-- Access log processes all requests for this virtual host.  By
            default, log files are created in the "logs" directory relative
to
            $CATALINA_HOME.  If you wish, you can specify a different
            directory with the "directory" attribute.  Specify either a
relative
            (to $CATALINA_HOME) or absolute path to the desired directory.
            This access log implementation is optimized for maximum
performance,
            but is hardcoded to support only the "common" and "combined"
patterns.
       -->
       <!--
       <Valve
className="org.apache.catalina.valves.FastCommonAccessLogValve"
                directory="logs"  prefix="localhost_access_log."
suffix=".txt"
                pattern="common" resolveHosts="false"/>
       -->
       <!-- Access log processes all requests for this virtual host.  By
            default, log files are created in the "logs" directory relative
to
            $CATALINA_HOME.  If you wish, you can specify a different
            directory with the "directory" attribute.  Specify either a
relative
            (to $CATALINA_HOME) or absolute path to the desired directory.
            This access log implementation is optimized for maximum
performance,
            but is hardcoded to support only the "common" and "combined"
patterns.

            This valve use NIO direct Byte Buffer to asynchornously store
the
            log.
       -->
       <!--
       <Valve
className="org.apache.catalina.valves.ByteBufferAccessLogValve"
                directory="logs"  prefix="localhost_access_log."
suffix=".txt"
                pattern="common" resolveHosts="false"/>
       -->

         <Resource name="mail/Session" auth="Container"
                   type="javax.mail.Session"/>
         <ResourceParams name="mail/Session">
           <parameter>
             <name>mail.smtp.host</name>
             <value>localhost</value>
           </parameter>
         </ResourceParams>

     </Host>
   </Engine>
 </Service>
</Server>

What I need to add/change in a code above ? All I find with google unclearly
describe the backend setup and describes only the code.  I'm a newbie with
servlets so I really can't wrote tomcat setup from a head . I need concrete
code.

p.s.
I didn't wont to be rude