The basic outline of putting static files into production consists of two\nsteps: run the collectstatic command when static files change, then\narrange for the collected static files directory (STATIC_ROOT) to be\nmoved to the static file server and served. Depending on\nSTATICFILES_STORAGE, files may need to be moved to a new location\nmanually or the post_process method of\nthe Storage class might take care of that.
Of course, as with all deployment tasks, the devil’s in the details. Every\nproduction setup will be a bit different, so you’ll need to adapt the basic\noutline to fit your needs. Below are a few common patterns that might help.
\nIf you want to serve your static files from the same server that’s already\nserving your site, the process may look something like:
\nPush your code up to the deployment server.
On the server, run collectstatic to copy all the static files\ninto STATIC_ROOT.
Configure your web server to serve the files in STATIC_ROOT\nunder the URL STATIC_URL. For example, here’s\nhow to do this with Apache and mod_wsgi.
You’ll probably want to automate this process, especially if you’ve got\nmultiple web servers.
\nMost larger Django sites use a separate Web server – i.e., one that’s not also\nrunning Django – for serving static files. This server often runs a different\ntype of web server – faster but less full-featured. Some common choices are:
\n\nConfiguring these servers is out of scope of this document; check each\nserver’s respective documentation for instructions.
\nSince your static file server won’t be running Django, you’ll need to modify\nthe deployment strategy to look something like:
\nWhen your static files change, run collectstatic locally.
Push your local STATIC_ROOT up to the static file server into the\ndirectory that’s being served. rsync is a\ncommon choice for this step since it only needs to transfer the bits of\nstatic files that have changed.
Another common tactic is to serve static files from a cloud storage provider\nlike Amazon’s S3 and/or a CDN (content delivery network). This lets you\nignore the problems of serving static files and can often make for\nfaster-loading Web pages (especially when using a CDN).
\nWhen using these services, the basic workflow would look a bit like the above,\nexcept that instead of using rsync to transfer your static files to the\nserver you’d need to transfer the static files to the storage provider or CDN.
There’s any number of ways you might do this, but if the provider has an API,\nyou can use a custom file storage backend\nto integrate the CDN with your Django project. If you’ve written or are using a\n3rd party custom storage backend, you can tell collectstatic to use\nit by setting STATICFILES_STORAGE to the storage engine.
For example, if you’ve written an S3 storage backend in\nmyproject.storage.S3Storage you could use it with:
STATICFILES_STORAGE = 'myproject.storage.S3Storage'\nOnce that’s done, all you have to do is run collectstatic and your\nstatic files would be pushed through your storage package up to S3. If you\nlater needed to switch to a different storage provider, you may only have to\nchange your STATICFILES_STORAGE setting.
For details on how you’d write one of these backends, see\nWriting a custom storage system. There are 3rd party apps available that\nprovide storage backends for many common file storage APIs. A good starting\npoint is the overview at djangopackages.org.
\nFor complete details on all the settings, commands, template tags, and other\npieces included in django.contrib.staticfiles, see the\nstaticfiles reference.