Connection Pooling Sql Server with Vaadin

I was trying to follow the notes from this forum post but I couldn't get it to work.

I've only just learned about connection pools so I have very little knowledge in setting it up, but I'm trying. Please don't get too mad.

We have a MS SQL server like this:

I want to access table1 in DB2

According to a tutorial, I'm suppose to go to Eclipse's Server > Server.xml and add the following code.

<Resource name="jdbc/ServerName"
        auth="Container"
        factory="org.apache.tomcat.dbcp.dbcp2.BasicDataSourceFactory"
        driverClassName="com.microsoft.sqlserver.jdbc.SQLServerDriver"
        type="javax.sql.DataSource"
        maxActive="50"
        maxIdle="10"
        maxWait="15000"
        username="username"
        password="password"
        url="jdbc:sqlserver://ServerName;instance=SQLEXPRESS;databaseName=DB2;"
        removeAbandoned="true"
        removeAbandonedTimeout="30"
        logAbandoned="true" />     

However, I read somewhere else it's actually suppose to be in Context.xml? Which I can't find in my Vaadin webapp. in This tutorial, it says it should be under the meta-inf folder but Every time I create a context.xml file there, it gets eaten when I  compile the program.

Next I add the drivers to my Tomcat library? I downloaded the microsoft ones from here

And I add this code to my java program:

Connection con = null;
        Statement stmt = null;
        ResultSet rs = null;
        try {
            Context ctx = (Context) new InitialContext().lookup("java:comp/env");
            BasicDataSource ds = (BasicDataSource)ctx.lookup("jdbc/ServerName");
            con = ds.getConnection();

            String sql = "select * from DB2.dbo.Table1 where [Field1] IS NULL";     
            stmt = con.createStatement();
            rs = stmt.executeQuery(sql);
            //Add in all info
            while (rs.next()){
                beanResultsList.addBean(new Bean(
                        rs.getInt(1),
                        rs.getString(2),
                        rs.getString(3),
                        rs.getDate(4),
                        rs.getDate(5),
                        rs.getInt(6),
                        rs.getInt(7),
                        rs.getInt(8),
                        rs.getInt(9),
                        rs.getString(10),
                        rs.getString(11),
                        rs.getString(12),
                        rs.getString(13)
                        ));
            }
            } catch (SQLException | NamingException e) {
                e.printStackTrace();
            }finally{
                try { con.close();  } catch (SQLException e) {}
                try { rs.close();   } catch (SQLException | NullPointerException e) {}
                try { stmt.close(); } catch (SQLException e) {}
            } // End finally (try catch)

And it doesn't work, I get nullpointerexception on the Con.close() because it's not reaching the sql server?

However, just using JTDS works fine, but it's really slow.

Connection con = null;
        Statement stmt = null;
        ResultSet rs = null;
        try {
            Class.forName("net.sourceforge.jtds.jdbc.Driver");
            con = DriverManager.getConnection("jdbc:jtds:sqlserver://ServerName;instance=SQLEXPRESS","username","password");

            String sql = "select * from DB2.dbo.Table1 where [Field1] IS NULL";     
                stmt = con.createStatement();
                rs = stmt.executeQuery(sql);
                //Add in all info
                while (rs.next()){
                    beanResultsList.addBean(new Bean(
                            rs.getInt(1),
                            rs.getString(2),
                            rs.getString(3),
                            rs.getDate(4),
                            rs.getDate(5),
                            rs.getInt(6),
                            rs.getInt(7),
                            rs.getInt(8),
                            rs.getInt(9),
                            rs.getString(10),
                            rs.getString(11),
                            rs.getString(12),
                            rs.getString(13)
                            ));
                }
                } catch (SQLException | NamingException e) {
                    e.printStackTrace();
                }finally{
                    try { con.close();  } catch (SQLException e) {}
                    try { rs.close();   } catch (SQLException | NullPointerException e) {}
                    try { stmt.close(); } catch (SQLException e) {}
                } // End finally (try catch)

Does anybody who has gotten connection pooling to work know what's wrong? And how do I get Context.xml in my vaadin app? And keep it there?

Ok well I managed to do this in the end with C3P0 and JTDS instead of using tomcat. I guess nobody knows how to do connection pooling on tomcat either so I'll post what I did here.

I don't think this is "best practice" but it worked for me really well. I'm not sure it will work for a larger environment though with heavy traffic.

In the UI, I just set up a ComboPooledDatasource and just passed them to each instance of the navigator.
It reduced the loadtime from 10-15 sec to 1-3 seconds! It also seems to reduce the memory it uses.

public class MyUI extends UI {
    private Navigator nav = new Navigator(this,this);
    private ComboPooledDataSource cpds;
    @Override
    protected void init(VaadinRequest vaadinRequest) {
        //Pooled connection
        cpds = new ComboPooledDataSource();
        try {
            cpds.setDriverClass("net.sourceforge.jtds.jdbc.Driver");
        } catch (PropertyVetoException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        } //loads the jdbc driver
        cpds.setJdbcUrl( "jdbc:jtds:sqlserver://ServerName/DB1" );
        cpds.setUser("username");
        cpds.setPassword("password");

        nav.addView(ViewOne.NAME,new ViewOne(cpds));
        nav.addView(ViewTwo.NAME, new ViewTwo(cpds));
}

Then to use it in the Views, I did this:

public class ViewTwo extends Panel implements View {

public ViewTwo(ComboPooledDataSource cpds){
}
        Connection con = null;
        PreparedStatement stmt = null;
        ResultSet rs = null;

        try {
            con =  cpds.getConnection();
            String sql = "select * from DB1";
            stmt = con.createStatement();
            rs = stmt.executeQuery(sql);
//...etc etc fill in BeanItemContainer for grid etc
}

But well, if anybody has any clue how to do it in Tomcat, I'd still like to know.