Rails app GET routes not working in production mode - ruby-on-rails

Question

I've inherited a Rails application that isn't completely functional when running in production mode. GET requests to the server are resulting in a no route match found error; however, when the server is run in development mode all routes will work, giving an expected 200 status.
Reviewing the code reveals that the application expects a prefixed subdomain in addition to the domain used in a successful URL request.
class ApplicationContext
def initialize(subdomain, host_with_port)
if subdomain.present?
case subdomain
when Rails.application.config.workninja.admin_subdomain
#environment = :admin
when Rails.application.config.workninja.mobile_subdomain
#environment = :mobile
else
#environment = :customer
end
else
raise 'Could not initialize ApplicationContext without subdomain'
end
#url_options = {subdomain: subdomain, host: host_with_port}
setup_method = "setup_#{#environment.to_s}_environment".to_sym
if respond_to?(setup_method, true)
send(setup_method, subdomain, host_with_port)
else
raise 'Unknown context environment'
end
end
attr_reader :environment
attr_reader :url_options
attr_reader :agency
def self.current
Thread.current['workninja:tenant_context']
end
def admin?
#environment == :admin
end
def mobile?
#environment == :mobile
end
def customer?
#environment == :customer
end
def ui_url_for(path, subdomain = url_options[:subdomain])
base = "#{Rails.application.config.workninja.https ? 'https' :
'http'}://#{subdomain}.#{Rails.application.config.workninja.domain}"
if Rails.application.config.workninja.html5mode
puts URI.join(base, path).to_s
else
puts URI.join(base, "/#/#{path}".gsub(/\/+/, '/')).to_s
end
end
The original front end supplied with the application constructs the request's URLs depending on the environment the sever was launched in.
{
"environment": "development",
"url": "http://admin.workninja.local:3000"
}
{
"environment": "production",
"url": "/api"
}
To me the production URL doesn't make sense as all it does is append "/api" to the root domain that the front end is hosted on. I can only assume that it's just a placeholder that needs to be replaced with the domain name that the rails server is hosted on once it's running in a live environment. The "/api" path isn't used throughout the functional development version of the app which makes me further assume it's a placeholder.
Using the above as a guide I replaced "/api" with "http://admin.workninja.com.au". After hosting the application on a live domain I confirmed it was working by running:
curl http://admin.workninja.com.com.au/auth -X POST
This gave me an expected error about credentials not being supplied but it shows that the server is actually receiving something. If you haven't realised the rails server when launched in production mode will respond to a POST request but still not a GET.
This is where my understanding of the problem breaks down. If
http://admin.workninja.local:3000/roles
works ("/roles being one of the applications routes") in a development environment why doesn't
http://admin.workninja.com.au/roles
work in a production environment as well? Can you assume from this fact that something isn't broken in the ruby codebase?
Below are some of the files relating to the configuration of the rails application in a production environment.
/config/deploy/production.rb
set :branch, 'master'
server 'ec2-54-66-230-174.ap-southeast-2.compute.amazonaws.com', user: 'ubuntu', roles: %w{app web db worker}
/config/environments/production.rb
Rails.application.configure do
# Settings specified here will take precedence over those in config/application.rb.
# Code is not reloaded between requests.
config.cache_classes = true
# Eager load code on boot. This eager loads most of Rails and
# your application in memory, allowing both threaded web servers
# and those relying on copy on write to perform better.
# Rake tasks automatically ignore this option for performance.
# This needs to be set to true in order for rails to launch in a production environment
config.eager_load = false
# Full error reports are disabled and caching is turned on.
config.consider_all_requests_local = false
config.action_controller.perform_caching = true
# Enable Rack::Cache to put a simple HTTP cache in front of your application
# Add `rack-cache` to your Gemfile before enabling this.
# For large-scale production use, consider using a caching reverse proxy like
# NGINX, varnish or squid.
# config.action_dispatch.rack_cache = true
# Disable serving static files from the `/public` folder by default since
# Apache or NGINX already handles this.
config.serve_static_files = ENV['RAILS_SERVE_STATIC_FILES'].present?
# Specifies the header that your server uses for sending files.
# config.action_dispatch.x_sendfile_header = 'X-Sendfile' # for Apache
# config.action_dispatch.x_sendfile_header = 'X-Accel-Redirect' # for NGINX
# Force all access to the app over SSL, use Strict-Transport-Security, and use secure cookies.
# config.force_ssl = true
# Use the lowest log level to ensure availability of diagnostic information
# when problems arise.
config.log_level = :warn
# Prepend all log lines with the following tags.
# config.log_tags = [ :subdomain, :uuid ]
# Use a different logger for distributed setups.
# config.logger = ActiveSupport::TaggedLogging.new(SyslogLogger.new)
# Use a different cache store in production.
# config.cache_store = :mem_cache_store
# Enable serving of images, stylesheets, and JavaScripts from an asset server.
# config.action_controller.asset_host = 'http://assets.example.com'
# Ignore bad email addresses and do not raise email delivery errors.
# Set this to true and configure the email server for immediate delivery to raise delivery errors.
# config.action_mailer.raise_delivery_errors = false
# Enable locale fallbacks for I18n (makes lookups for any locale fall back to
# the I18n.default_locale when a translation cannot be found).
config.i18n.fallbacks = true
# Send deprecation notices to registered listeners.
config.active_support.deprecation = :notify
# Use default logging formatter so that PID and timestamp are not suppressed.
config.log_formatter = ::Logger::Formatter.new
# Do not dump schema after migrations.
config.active_record.dump_schema_after_migration = false
# Application hostname
config.surgeforce.domain = 'surgeforce.com.au'
config.surgeforce.https = false
config.surgeforce.html5mode = true
end
/config/puma.rb
threads 1, 6
workers Integer(ENV['PUMA_WORKERS'] || 3)
on_worker_boot do
require "active_record"
cwd = File.dirname(__FILE__)+"/.."
ActiveRecord::Base.connection.disconnect! rescue ActiveRecord::ConnectionNotEstablished
ActiveRecord::Base.establish_connection(ENV["DATABASE_URL"] || YAML.load_file("#{cwd}/config/database.yml")[ENV["RAILS_ENV"]])
end
If you believe any other piece of the applications code to be critical in the investigation let me know and I'll include it.

Answer 1

The problem is that Rails' subdomain method is very simple and doesn't know anything about the structure of com.au domains. For "admin.workninja.com.au" which you use on production, the subdomain method would return "admin.workninja". From the docs:
Returns all the \subdomains as a string, so "dev.www" would be
returned for "dev.www.rubyonrails.org". You can specify a different tld_length, such as 2 to catch "www" instead of "www.rubyonrails" in "www.rubyonrails.co.uk".
And – without knowing your configuration– "admin.workninja" will very likely not match your config.workninja.admin_subdomain configuration anymore.
The solution is to configure a tld length of 2 on production. Just add the following to the configuration block in your config/environments/production.rb:
config.action_dispatch.tld_length = 2 # Defaults to 1