Trouble with rails multistage deployment on Ubuntu 11.10

Question

I am trying to get a multistage deployment on my server and I am having trouble getting a staging and production configuration running.

Server: Ubuntu 11.10 (GNU/Linux 3.0.0-19-server x86_64)

Here is what I have done so far:

1. I created a new project called form_tester on my local machine. I am using a simple database with sqlite3 locally. I wish to use mysql2 on both staging and production.

2. Setup directories for staging and production on server.

   /home/username/form_tester/staging
   /home/username/form_tester/production
   

3. Made two databases on server for staging and production. I also configured them for access by a user with the correct privileges since I am able to generate the tables in them.

   ftstaging
   ftproduction
   

4. Configured the config/database.yml file for both staging and production.

   staging: 
     adapter: mysql2 
     encoding: utf8 
     reconnect: false
     database: ftstaging
     pool: 5 
     username: dbuser 
     password: pass 
     host: server.com
   
   production: 
     adapter: mysql2 
     encoding: utf8 
     reconnect: false
     database: ftproduction
     pool: 5 
     username: dbusr 
     password: pass 
     host: server.com
   

5. Configured the Gemfile for both staging and production.

   source 'https://rubygems.org'

   gem 'rails', '3.2.8'
   gem 'jquery-rails'
   gem 'capistrano'
   
   group :development do
     gem 'sqlite3'
   end
   
   group :staging do
     gem 'activerecord-mysql2-adapter'
     gem 'mysql2'
   end
   
   group :production do
     gem 'activerecord-mysql2-adapter'
     gem 'mysql2'
   end
   
   # Gems used only for assets and not required
   # in production environments by default.
   group :assets do
     gem 'sass-rails',   '~> 3.2.3'
     gem 'coffee-rails', '~> 3.2.1'
     gem 'uglifier', '>= 1.0.3'
   end
   

6. Ran Capify to get deployment files. Deleted everything from the deploy.rb file. Left the Capfile alone.

7. Created deploy directory and made two files.

   config/deploy/production.rb
   config/deploy/staging.rb
   

8. Added the following details for configuration:

For config/deploy/production.rb:

    server "server.com", :app, :web, :db, :primary => true
    set :deploy_to, "/home/username/form_tester/production"
    set :rails_env, "production"

For config/deploy/staging.rb:

    server "server.com", :app, :web, :db, :primary => true
    set :deploy_to, "/home/username/form_tester/staging"
    set :rails_env, "staging"

For config/deploy.rb:

    set :stages, ['production', 'staging']
    set :default_stage, 'staging'
    require 'capistrano/ext/multistage'

    # Set application name
    set :application, 'form_tester'
    set :domain, 'server.com'
    set :user, 'username'

    # Use Git source control
    set :scm, :git
    set :repository, "ssh://#{user}@#{domain}/home/#{user}/git/#{application}.git"
    set :branch, 'master'
    set :deploy_via, :remote_cache
    set :scm_verbose, true

    default_run_options[:pty] = true
    set :use_sudo, false

    namespace :deploy do
      task :start do ; end
      task :stop do ; end

      desc "Restart application"
      task :restart, :roles => :app, :except => { :no_release => true } do
        run "#{try_sudo} touch #{File.join(current_path,'tmp','restart.txt')}"
      end
    end

9.Initialized git repo on server with git --bare init and made sure I have proper access to the directory for the current username.

10.Initialized git repo on dev machine and pushed the files to the server repo at
ssh://username@server.com/home/username/git/form_tester.git.

11.Begin deployment from local machine to staging on server. I run all the following commands with no errors.

To build the directory tree under /home/username/form_tester/staging

    $ cap deploy:setup
    $ cap deploy:check
    $ cap deploy:update

    *ssh'ed into my server*

    $ rake schema:db:load
    $ rake db:seed
    $ rails console
    > app.get('/')
    I get a '200' returned
    > exit

still ssh'ed into my server

12.I made a symlink from /var/www/staging to my application public directory

    /home/username/form_tester/staging/current/public.

    $ sudo ln -s /home/username/form_tester/staging/current/public /var/www/staging

13.Modified the /etc/apache2/sites-available/default file to add a subURI for the staging app.

    $ sudo vi /etc/apache2/sites-available/default

    …
            RailsBaseURI /staging
            <Directory /var/www/staging>
                    Options -MultiViews
            </Directory>
    …

14.At this point (locally) I run cap deploy:restart and it seems to restart the server since there is no error. (logged into the server) I also tried sudo service apache2 restart, which also restarts the server fine.

15.At this point I pull up the url server.com/staging but don't see my app.

Other files:

config/application.rb:

    require File.expand_path('../boot', __FILE__)

    require 'rails/all'

    if defined?(Bundler)
      Bundler.require(*Rails.groups(:assets => %w(development test)))
    end

    module FormTester
      class Application < Rails::Application
        config.encoding = "utf-8"
        config.filter_parameters += [:password]
        config.active_support.escape_html_entities_in_json = true   
        config.active_record.whitelist_attributes = true
        config.assets.enabled = true
        config.assets.version = '1.0'
      end
    end

config/environments/staging.rb and production.rb

    FormTester::Application.configure do
      config.cache_classes = true
      config.consider_all_requests_local       = false
      config.action_controller.perform_caching = true
      config.serve_static_assets = false
      config.assets.compress = true
      config.assets.compile = false
      config.assets.digest = true
      config.i18n.fallbacks = true
      config.active_support.deprecation = :notify
    end

/var/log/apache2/error.log

There seems to be nothing ruby related showing up here except phusion passenger, but it's not an error.

    [Thu Nov 01 01:16:11 2012] [notice] caught SIGTERM, shutting down
    PHP Warning:  PHP Startup: Unable to load dynamic library '/usr/lib/php5/20090626/gd.so' - /usr/lib/php5/20090626/gd.so: cannot open shared object file: No such file or directory in Unknown on line 0
    PHP Warning:  PHP Startup: Unable to load dynamic library '/usr/lib/php5/20090626/mcrypt.so' - /usr/lib/php5/20090626/mcrypt.so: cannot open shared object file: No such file or directory in Unknown on line 0
    [Thu Nov 01 01:16:12 2012] [notice] Apache/2.2.20 (Ubuntu) DAV/2 SVN/1.6.12 PHP/5.3.6-13ubuntu3.7 with Suhosin-Patch Phusion_Passenger/3.0.17 configured -- resuming normal operations
    [Thu Nov 01 01:16:20 2012] [error] [client 68.6.38.254] File does not exist: /var/www/favicon.ico
    [Thu Nov 01 01:16:22 2012] [error] [client 68.6.38.254] File does not exist: /var/www/favicon.ico
    [Thu Nov 01 01:23:07 2012] [error] [client 68.6.38.254] File does not exist: /var/www/favicon.ico

Observations:

1. Running rake schema:db:load appears to build the tables properly for my database but its actually building them for the development.sqlite3 file.

2. If I open the rails console and run Rails.env, I get "staging" as the environment.

3. This is most likely tied with #1, rake db:seed doesn't seem to fill my database called "ftstaging" in mysql on the server even though it executes.

4. cap deploy doesn't seem to do anything different than cap deploy:update.

5. The end result is a page at server.com/staging that says "We're sorry, but something went wrong."

6. There doesn't seem to be anything shown in my

7. Running rake about shows that I am still in development mode.

   About your application's environment
   Ruby version              1.9.3 (x86_64-linux)
   RubyGems version          1.8.24
   Rack version              1.4
   Rails version             3.2.8
   JavaScript Runtime        Node.js (V8)
   Active Record version     3.2.8
   Action Pack version       3.2.8
   Active Resource version   3.2.8
   Action Mailer version     3.2.8
   Active Support version    3.2.8
   Middleware                ActionDispatch::Static, Rack::Lock, #<ActiveSupport::Cache::Strategy::LocalCache::Middleware:0x00000002796948>, Rack::Runtime, Rack::MethodOverride, ActionDispatch::RequestId, Rails::Rack::Logger, ActionDispatch::ShowExceptions, ActionDispatch::DebugExceptions, ActionDispatch::RemoteIp, ActionDispatch::Reloader, ActionDispatch::Callbacks, ActiveRecord::ConnectionAdapters::ConnectionManagement, ActiveRecord::QueryCache, ActionDispatch::Cookies, ActionDispatch::Session::CookieStore, ActionDispatch::Flash, ActionDispatch::ParamsParser, ActionDispatch::Head, Rack::ConditionalGet, Rack::ETag, ActionDispatch::BestStandardsSupport
   Application root          /home/username/form_tester/staging/releases/20121101080752
   Environment               development
   Database adapter          sqlite3
   Database schema version   20121030011807
   

I have tried the following tutorials with little luck:

• https://github.com/capistrano/capistrano/wiki/2.x-Multistage-Extension
• http://trevmex.com/post/367574671/steps-for-setting-up-a-capistrano-multistage-deployment
• http://cjohansen.no/en/apache/multi_staging_environment_for_rails_using_capistrano_and_mod_rails
• http://guides.beanstalkapp.com/deployments/deploy-with-capistrano.html#prepare-project

Questions:

I'm sure there is someone out there that understands the deployment process in and out, what am I missing? What should I check for? From what I have shown here, what can I do differently?

Any help is greatly appreciated at this point. I have been fighting with this issue for two days now and don't know what I'm doing wrong and don't know what else to try. I just started using rails about 3 weeks ago and most of my frustration has been due to not being able to find a clear description of the process for deployment. I hope to gain clarity on this issue and in return help others get from idea to multistage deployment easily.

On another note, why would someone "down vote" this post? I'm asking a legit question with information to back it up. If it's not a good question please tell me why. Thanks

---

11/1/12 - UPDATE:

I figured out how to get my app into the "staging" environment but it still doesn't load at server.com/staging.

I added "Rails.env = ActiveSupport::StringInquirer.new('staging')" into my config/application.rb file, but I don't know why this forces the environment. Is this the correct way to do it?

I then deployed, ssh'ed into my server as username, and ran rake db:migrate and rake db:seed. I now see my database populated.

What else could be going on? Is there something wrong with my environments/staging.rb file?

---

/etc/apache2/httpd.conf

ServerName server.com

LoadModule passenger_module /usr/local/rvm/gems/ruby-1.9.3-p194/gems/passenger-3.0.17/ext/apache2/mod_passenger.so
PassengerRoot /usr/local/rvm/gems/ruby-1.9.3-p194/gems/passenger-3.0.17
PassengerRuby /usr/local/rvm/wrappers/ruby-1.9.3-p194/ruby

Answer 1

I FINALLY discovered the the primary issue: the asset pipeline

First issue: I didn't uncomment the following line form the Capfile

load 'deploy/assets'

AND

Second issue: I never compiled the assets (on development machine or server side. big whoops!)

$ bundle exec rake assets:precompile

References

• http://guides.rubyonrails.org/asset_pipeline.html
• Rails/Bundler precompile vs lazy compile

Followup question to anyone who may know: I noticed that when I compiled the assets, it says RAILS_ENV=production instead of staging. When making the staging.rb file I basically copied the production.rb file. I'm not sure if that has anything to do with it saying production.

Is this expected?

/usr/local/rvm/rubies/ruby-1.9.3-p194/bin/ruby /usr/local/rvm/gems/ruby-1.9.3-p194@global/bin/rake assets:precompile:all RAILS_ENV=production RAILS_GROUPS=assets