0

Possible Duplicate:
Can you help me with my capacity planning?

I am looking for advice for our web server configuration.

  • SaaS model startup (same app, different databases and subdomains).

  • Serving 50+ sites right now.

  • Dedicated server with 8GB of Memory

  • Only about 20% of them have a moderate traffic.

  • Our configuration is Nginx on port 80 with a reverse proxy to apache where python/django is run via mod_wsgi, Centos is the Linux Distribution.

We are looking for alternatives because each low (even if by low we mean none) traffic site consumes about 18 Mb, we have already memory warnings. So what advices do you have?.

Will serving directly from nginx with gunicorn/fastcgi/uWSGI increase our capacity to serve many more sites? what other kind of confiuguration do yo reccomend?.

I have a mostly developer background so any explanation would be welcomed. Thanks....

Community
  • 1
quarry32
  • 255
  • 1
  • 11

1 Answers1

0

I would diwtch mod_wsgi and switch to uwsgi. Run it in emperor mode so you have 1 uwsgi process for all vhosts

http://uwsgi-docs.readthedocs.org/en/latest/Emperor.html

Mike
  • 21,910
  • 7
  • 55
  • 79