I have a Rails app installed on a Slicehost server running Apache 2 and Ubuntu LTC 10.04. Things have worked beautifully up until now: I edit a file, do a quick mongrel_rails cluster::restart
, and the changes are reflected in production. However, suddenly this process has broken down.
For example, I have a class called Master
located in /lib/master.rb
. I added a new method to this class that simply runs puts "it works!"
, then restarted the mongrel cluster. Looking at the production logs, the server throws an error and thinks this method doesn't exist. When I go to the console using ruby script/console production
, however, I can use this new method perfectly. I even tried deleting the file containing entire Master
class. Once again, the production thought it was still there, but the production console correctly recognized it was missing.
Any ideas? How can the production environment detect a class that doesn't even exist anymore?