How to configure database.yml for deployment to Heroku

Question

I recently upgraded to the newest version of Rails, and I don't understand how to deploy applications to Heroku.

Here is my database.yml file

default: &default
  adapter: postgresql
  pool: 5
  timeout: 5000

development:
  <<: *default
  database: db/development.sqlite3

test:
  <<: *default
  database: db/test.sqlite3

production:
  <<: *default
  database: db/production.sqlite3

I have never seen this syntax before in database.yml. Does anyone know how to configure this?

It looks a lot different than what I'm used to

development:
 adapter: mysql2
 encoding: utf8
 database: my_app_development
 pool: 5
 username: root
 password:

test:
 adapter: mysql2
 encoding: utf8
 database: my_app_test
 pool: 5
 username: root
 password:


production:
 adapter: mysql2
 encoding: utf8
 database: ymca_gym_production
 pool: 5
 username: root
 password:

Thanks

Answer 1

I don't know if this is exactly what you want, but I just spent a while figuring this out myself, so I thought I'd post it here.

ActiveRecord supports getting database connection information both from the environmental variable DATABASE_URL and from the database.yml file. Exactly which one gets used in any situation can be a little tricky. Fortunately, this has already mostly been taken care of - see this pull request for most of the details.

The short version is that if there isn't a url key in your database.yml, then the information in DATABASE_URL automatically overrides database.yml for everything that it can set that is present in the URL, including username, password, server, port, adapter, and database name. If there is a url key in database.yml, then that overrides the DATABASE_URL env var and everything else that it conflicts with in the database.yml entry.

Since Heroku sets a full DATABASE_URL automatically, basically everything except for the pool and encoding are taken from that, and entries for the other things in the production section of the database.yml are ignored. So you can leave them blank, or set anything that is more convenient for you on your dev machine.

Since it is possible to put ERB in yml files, some sources recommend that you explicitly put

production:
  url: <%= ENV['DATABASE_URL'] %>

in your database.yml to explicitly set the url as the env var. This makes it more clear that whatever is in the env var is meant to override everything in the yml.

Answer 2

For Heroku you will have to use postgresql as it doesn't support mysql2. Heroku has its own mechanism to handle databases which you can read more about here: https://devcenter.heroku.com/articles/heroku-postgresql

Essentially, treat "heroku's databases" and local databases that you define in this file completely different. It will be easier for you to use sqlite for local and test environments and for production you should change your yaml code to this :

development:
 adapter: mysql2
 encoding: utf8
 database: my_app_development
 pool: 5
 username: root
 password:

test:
  adapter: sqlite3
  database: db/development.sqlite3
  pool: 5
  timeout: 5000

production:
      adapter: postgresql
      database: my_database_production
      pool: 5
      timeout: 5000

Above code is not enough to get it working on heroku yet, you will also need to edit the gemfile content like below :

gem 'pg', :group => :production
gem 'mysql2' , :group => :development
gem 'sqlite3', :group => :test

I have made the gemfile code according to the database.yaml code that i wrote. You can use mysql2 for both development and test environment. If you are doing so you can change the gemfile contents like below:

gem 'pg', :group => :production
gem 'mysql2' , :group => [:development, :test]

Hope this helps.. :)