Question

I am developing a small part of a PHP application with some python code. The python code runs like an equivalent of a servlet (listens and responds to HTTP on port 8765) on localhost. The PHP app calls it like:

PHP'S_CURL("http://localhost:8765/search?term=electrical+design")

The pyth-let is written with the BaseHTTPServer module like:

class MyHandler(BaseHTTPRequestHandler):
   def do_GET(self):
      if self.path=="/search":
         self.send_response(200)

         # ....

         self.wfile.write(st)

It works on my workstation and my colleague’s. I now want to deploy it in a production environment, with modifications. The idea in mind is that I should:

1. modify my app to FCGI
2. get an inexpensive VPS account
3. set Apache to use FCGI to spawn and keep alive both the PHP app and the pyth-let.

So it’s a localhost app, which shouldnt be exposed publically. There should be reliable way to keep it alive. We expect ~800 hits a day before needing an upgrade, so only a single instance need be kept alive.

Is there a feasible way to do this on a popular shared host, rather than a VPS? Am I on the right track with my above-mentioned plan?

---

Postscipts
I mentioned “easy, reliable, cheap way”
and by “way” I meant both the development direction as well as a good (cheap) hosting plan that can support it.

Answer 1

You can use flup to serve your python app with fastcgi. I’ve also used gunicorn to deploy python webapps along with supervisor and found that to be a good approach – and even easier to setup.