Deployment

In order to run a Sprinkles site on a production system, you have several options.

Using The Built-In Warp Server

This is what you’ve been doing in the Getting Started guide: Sprinkles has a web server built into it that you can run as a standalone process.

For a production server, however, you will want some extra bells and whistles, such as running the application on port 80 (which is a privileged port and thus requires root permissions on most systems), serving over HTTPS (which Sprinkles itself does not implement), running on a subdomain, etc. Existing web servers like nginx or Apache are excellent for this, and a simple setup is to configure one of them as a Reverse Proxy, forwarding requests to a Sprinkles process that listens on a non-privileged port.

A bit of extra work is needed to keep this working reliably, but once everything is in place, this is a very flexible and performant setup.

FastCGI

With FastCGI, the main web server process fires up one or more Sprinkles processes as needed, and keeps them running for subsequent requests. For this to work well, you’ll want to create a wrapper script like the following (put it right next to your project.yml), and point your web server’s FastCGI handler to it:

#!/bin/sh
sprinkles -fcgi

SCGI

SCGI is similar to FastCGI; the protocol is different, but the use cases are mostly the same. A suitable wrapper script looks like this:

#!/bin/sh
sprinkles -scgi

CGI

CGI is not your best choice, because it requires going through all the startup routines for every request; however, if you must, Sprinkles can work as a CGI script. Just like with FastCGI, you probably want a little wrapper to point your web server to:

#!/bin/sh
sprinkles -cgi