Skip to content

Latest commit

 

History

History
104 lines (73 loc) · 3.45 KB

demo.md

File metadata and controls

104 lines (73 loc) · 3.45 KB
Raw

Week 2 demo: simple web server

Create the directory named demo:

mkdir demo
cd demo

Create the file hello.html in this directory:

<html>
    <head>
        <title>Demo demo demo!</title>
    </head>
    <body>
        <h2>Hello from webserver!</h2>
    </body>
</html>

Start the web server in this directory:

python3 -m http.server

Note that the web server is started on port 8000.

Request the web resource from the browser: http://localhost:8000/hello.html

URL parts:

  • http: protocol
  • locahost:8000: web server address and port
  • hello.html: requested resource

Note that the resource (HTML document created above) is fetched by the client from the web server and transformed (rendered) to a human readable text. This is possible because:

  • Web client knows how to talk HTTP to the web server and retrieve documents
  • Web server understands HTTP
  • Web client understands HTML and knows how to render it

Send another request to the web server without specifying any exact resource: http://localhost:8000.

Note that if no resource is requested, web server just lists all available resources in this directory. This list is called directory index. This behavior is common convention for most web servers.

Go back to the terminal where web server logs are printed. You will see something similar to:

Serving HTTP on 0.0.0.0 port 8000 ...
127.0.0.1 - - [17/Sep/2019 11:47:30] "GET /hello.html HTTP/1.1" 200 -
127.0.0.1 - - [17/Sep/2019 11:49:41] "GET / HTTP/1.1" 200 -

Web server is telling about the requests it got, and how did it respond:

  • GET means that the client has requested some resource
  • /hello.html is the requested path, / means no resource were requested
  • HTTP/1.1 is the protocol the client used for request
  • 200 is the server response code (means that resource was found on server and sent to client)

Web browser is not the only possible web client; there are command-line clients, clients being part of other services etc.

Try some command line web client:

wget --output-document=- http://localhost:8000/hello.html
wget --output-document=- http://localhost:8000
curl http://localhost:8000/hello.html
curl http://localhost:8000

You may need to install cURL first, on Debian/Ubuntu/etc. run sudo apt install curl the last two commands are not working.

Note that directory listing is also a HTML page. This is created by web server automatically.

Hint: start using wget, curl or better, both. Get some practice with them. These are essential tools and every developer, tester, devops and sysadmin must know how to use them.

HTTP is text-based protocol, and it can be easily emulated manually. You don't even need web client. Run the telnet session and connect to the web server:

telnet localhost 8000

You may need to install telnet first, on Debian/Ubuntu/etc. run sudo apt install telnet the command above is not working.

Request a resource from the web server:

GET /hello.html

Note that we didn't specify the protocol. Web server is smart enough to choose one automatically.

This example with telnet is purely artificial though. You'll rarely need telnet to talk to web servers; web browsers and command line tools mentioned above (wget and curl) should cover most of the needs.

You should now have the basic understanding how web clients talk to web servers using HTTP protocol.

Make sure to stop the web server (Ctrl+C) once done with your experiments!