Skip to main content

django-wsgiserver installs a web server for Django using CherryPy's WSGI server.

Project description


django-wsgiserver is a django app for serving Django sites via CherryPy’s excellent, production ready, pure-python web server without needing to install all of Cherrypy. Note that Cherrypy names its server “wsgiserver” but it is actually a full-blown, multi-threaded http/https web server which has been used on production sites alone, or more often proxied behind a something like Apache or nginx.


The wsgiserver component has been used for years in production. Peter Baumgartner noted that it solved problems for him on memory on a memory-limited VPS hosted site [1]. Performance-wise it does well: it can serve up thousands of requests per second [Piel2010].

I haven’t used django-wsgiserver for production myself (yet) as daemonized modwsgi and uwsgi have served me well. I use it all the time though during development. It’s my “pocket-sized” server. completely written in python and it gives me an instant approximation of the final production environment I use. In some ways, it’s much better than the development server which is built into django. It’s noticeable when I have pages that do multiple ajax calls and the built-in runserver hangs. I just stop the built-in server, and then do

$ runwsgiserver

and reload my browser page and the problem is fixed. It’s also useful to see if some weird effect is being caused by runserver’s process of loading the settings twice.

This project is a slight modification of code form Peter Baumgartner code (see Peter’s blog post) Peter and others did the work of creating a management command. I’ve added a few small improvements from my point of view: It doesn’t require installing cherrypy separately. It uses the same port as the development server (8000) so I don’t need to re-enter my testing url in my browser, and it works by default with OS’s like Mac OS 10.6 and Ubuntu 10.04 which prefer binding localhost to an ip6 address.


To get started using the server, you need nothing outside of django itself and the project code that you would like to serve up. However, for ssl support, you may need PyOpenSSL–though the new cherrypy server includes support for using the python built-in ssl module depending on which version of python you are using.


To install, django-wsgiserver follows the usual pattern for a django python application. You have several options

  1. pip install django-wsgiserver OR
  2. easy_install django-wsgiserver
  3. If you would like to use the latest possible version, you can use pip and mercurial checkout from bitbucket
pip install -e hg+
  1. Alternatively you can download the code and install it so that django_wsgiserver is on your PYTHONPATH

After you used one of the methods above, you need to add django_wsgiserver to your INSTALLED_APPS in your django project’s settings file


To see how to run the server as a management command, run:

$ python runwsgiserver help

from within your project directory. You’ll see something like what’s below:

Run this project in CherryPy's production quality http webserver.
Note that it's called wsgiserver but it is actually a complete http server.

  runwsgiserver [options] [wsgi settings] [stop]

  Optional CherryPy server settings: (setting=value)
    host=HOSTNAME         hostname to listen on
                          Defaults to,
                          (set to to bind all ip4 interfaces or :: for
                          all ip6 interfaces)
    port=PORTNUM          port to listen on
                          Defaults to 8000
    server_name=STRING    CherryPy's SERVER_NAME environ entry
                          Defaults to localhost
    daemonize=BOOL        whether to detach from terminal
                          Defaults to False
    pidfile=FILE          write the spawned process-id to this file
    workdir=DIRECTORY     change to this directory when daemonizing
    threads=NUMBER        Number of threads for server to use
    ssl_certificate=FILE  SSL certificate file
    ssl_private_key=FILE  SSL private key file
    server_user=STRING    user to run daemonized process
                          Defaults to www-data
    server_group=STRING   group to daemonized process
                          Defaults to www-data

    Run a "standard" CherryPy server server
      $ runwsgiserver

    Run a CherryPy server on port 80
  $ runwsgiserver port=80

Run a CherryPy server as a daemon and write the spawned PID in a file
  $ runwsgiserver daemonize=true pidfile=/var/run/

    Run a CherryPy server using ssl with test certificates located in /tmp
  $ runwsgiserver ssl_certificate=/tmp/testserver.crt ssl_private_key=/tmp/testserver.key


If you want to use an installed version of Cherrypy–perhaps because you now have a more recent version, you only need to change one line of code in (around line 177) of django_wsgiserver/management/commands/

from django_wsgiserver.wsgiserver import CherryPyWSGIServer as Server
#from cherrypy.wsgiserver import CherryPyWSGIServer as Server

Just comment out the import from django_wsgiserver.wsgiserver and uncomment the import from cherrypy.wsgiserver to make the switch.

To do

  • looking at settings for serving static media automatically?
  • get in touch with Peter/see if a merge would be desirable
  • I should probably just add a switch to allow use of the native cherrypy install
  • upload to the cheeseshop/pypi at some point.


  • added test project in tests/ directory
  • got tired of typing run_cp_wsgiserver so did a rename so I could use runwsgiserver instead.
  • updated wsgiserver to svn r2680 (matches cherrypy version 3.2 beta+). This fixes some bugs and gives better python 2.6 support. This version of cherrypy will also support python 3.x for whenever django starts supporting it.
  • use port 8000 as with django devserver rather than Cherrypy’s default 8088
  • adapted defaults host= in order to work with ip4 by default. This addresses an issue I first noticed on mac OS 10.6 and later on ubuntu 10.04 where ip6 is active by default. Can get around this by adjusting the host option. For binding all ip4 interfaces, set to For all ip6 interfaces I believe you use ‘::’ You can also bind a specific interface by specifying host=<specific ip address> See
  • switched code to use run_cp_wsgiserver instead of runcpserver

Acknowledgments and References

Many thanks to Peter and lincolnloop for describing how to do this and writing the code.

Peter acknowledged idea and code snippets borrowed from Loic d’Anterroches, adapted to run as a management command

Note, there is also similar code on PyPi and at maintained by Alexander Solovyov

The latest version of the cherrypy wsgiserver can be retrieved with:

svn co

Peter hosts his code at

[1]For example Peter’s blog post describes using django_cpserver on a VPS.

Project details

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page