Before joining 37signals I hadn't encountered Puma-dev. This post is a message to past-me who would have benefited from using it to develop a constellation of Ruby on Rails applications.
A pretty typical setup for an organisation with a Ruby on Rails monolith is to at some point start breaking it up. Examples of Rails apps that may be extracted out:
A pretty typical setup for an organisation with a Ruby on Rails monolith is to at some point start breaking it up. Examples of Rails apps that may be extracted out:
- A specialised search or newsfeed API
- Authentication services
- Admin UI
- Microservices
Working on a system like this locally requires that the developer coordinates running multiple Rails servers:
cd ~/projects/org/web rails server -p 3000 # another terminal window cd ~/projects/org/authentication rails server -p 3001
Then, to call an API for example, each application needs to be configured with URLs to point to the other, with URLs like http://localhost:3001
Puma-dev presents a better way
By running a background process to manage all the Rails apps, starting them, and shutting them down when idle, we can avoid the tedium of managing multiple Rails servers in multiple terminal windows. We also get:
- HTTPS support
- .test TLD
So instead of http://localhost:3000 and http://localhost:3001 we get https://web.test and https://auth.test (or whatever name we like).
brew install puma/puma/puma-dev # Configure some DNS settings that have to be done as root sudo puma-dev -setup # Configure puma-dev to run in the background on ports 80 and 443 with the domain `.test`. puma-dev -install
# Gemfile gem "puma"
Step 2
In each application you'd like to use with Puma-dev, symlink the application directory into ~/.puma-dev. The puma-dev link command can help do that:
cd ~/projects/org/authentication puma-dev link -n auth
Step 3
In development ensure that Rails is configured to allow requests from any host.
# config/environments/development.rb Rails.application.configure do ... config.hosts = [] ... end
If you run into problems you can check the Puma-dev logs:
tail -f ~/Library/Logs/puma-dev.log
You might need to restart Puma-dev after you make a configuration change:
puma-dev --stop
Consistent URLs
If you're working as part of a team, it's useful to know the domain that other developers will be using to develop the application locally.
To share a consistent URL, rather than run the puma-dev link command manually, you can add it to your Rails bin/setup file.
# bin/setup # Existing Rails setup ... puts "\n== Configuring Puma-dev to host https://auth.test ==" system! "puma-dev link -n auth"
That way you establish a convention for all developers on the team. Run bin/setup and get a consistent development environment with a consistent URL.
Debugging
With your Raills app running the background you might wonder how you can use a debugger such as byebug or Ruby debug.
Thanks to remote debugging, this is possible. Here's how you can configure your Rails app to support remote debugging with Ruby debug.
Step 1
Add the Ruby debug gem to your Gemfile.
group :development, :test do gem "debug", require: false end
In your test environment the standard breaking debugger will be useful:
# config/environments/test.rb require "debug" Rails.application.configure do ...
But in the development environment you'll need to use open_nonstop:
# config/environments/development.rb require "debug/open_nonstop" Rails.application.configure do ...
This way, when you insert a breakpoint with debugger in development, you'll be able to remotely attach to the debugger:
rdbg --attach
Conclusion
With this Puma-dev setup it should be less tedious to run multiple Rails app locally while maintaining the benefits of debugging.