The goal of this article is to create a "Production-Quality" Rails Server. Thus, the best possible technologies (strictly my opinion) have been selected at the time of this writing to achieve this goal.

Unlike other developers, I prefer to install my Rails Applications under Web Subdirectories, such as http://www.not404.com/MyRailsApp, instead of running it as a Root Application of a Web Root, such as http://MyRailsApp.not404.com/. These instructions are geared for how I lay things out, but will let you know what to adjust in order to run your Rails Apps as traditional Web-Root Applications.

You may also notice that these instructions are SQLite3-oriented. This is intentional. IMHO, it's better to use the simplest-case database to prove that everything else is properly stitched together. Then, once you're satisfied that everything is properly locked down and performance-tuned, you can focus your attention on tying your Rails Application to a real database.

• Fedora Core (6 or later)
• Ruby on Rails
• Mongrel Ruby Application Server
• Lighttpd Web Server
• Various Plumbing and Configuration "Best Practices" to stitch it all together. :-)

Related Articles:

• How To Connect To SQL2005 From Ruby On Rails on Fedora Core 6 (Linux)

I've been very fortunate to lead an ETL project, using Ruby on Rails as the core engine. With a few weeks devoted to "Pure R&D Effort", I have come up with a mini-recipe of tweaks and connectivity layers necessary to connect a Rails Application (on Linux) to a SQL2005 Database.  I hope my published instructions here are eloquent and sufficient enough to guide you through your own efforts along such a path.

System Requirements and Configuration Steps Involved

• UnixODBC >= 2.2.11
• FreeTDS >= 0.63
• Configure an ODBC DSN
• Ruby ODBC >= 0.996
• Rails ODBC >= 1.1
• Ruby DBI >= 0.1.1
• Various Plumbing and Best-Practices