NYULibraries Deploy

NYULibraries Deploy offers a set of common rake tasks for the NYU libraries, as well as a set of common capistrano recipes mixed into a delicious default recipe. Most likely the capistrano recipes will handle most of the Rake tasks, but they can be used independently.

Rake Tasks

New Relic

Since the NYU Libraries uses the rails_config gem, we need a task to explicitly load the relevant settings (e.g. license key, app name) to newrelic.yml in the absence of a Rails environment.

rake nyu:newrelic:set
rake nyu:newrelic:reset

Puma

Start, stop, and restart tasks for the Puma webserver.

rake nyu:puma:start[9292]
rake nyu:puma:restart[9292]
rake nyu:puma:stop[9292]

Capistrano Recipes

Assets

This provides recipes that allow us to precompile the assets (only if their were changes!) and to create a symlink for the assets.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/assets'

Easy, the asset precompilation will work automatically after the capistrano's deploy:update_code task, and the symlinking will work automatically before capistrano's finalize_update.

If you need, you can always run cap deploy:assets:precompile or cap deploy:assets:symlink in a CLI.

Bundler

All this does is that it loads up Bundler's capistrano recipe and hooks deploy:migrate to run after it.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/bundler'

Cache

Clears rail cache!

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/cache'

Or you can run cap cache:tmp_clear.

Multistage

Loads up capistrano's multistage. Thats it.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/multistage'
# Same as
require 'capistrano/ext/multistage

New Relic

Loads up New Relic's capistrano recipes and adds two new tasks, newrelic:set and newrelic:reset.

cap newrelic:set just runs the rake task defined above, and cap newrelic:set runs the reset one. Also hooks New Relic's newrelic:notice_deployment to run after deploy:update.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/new_relic'

Rails Config

This beast makes use of the rails_config gem and loads the relevant settings (app name, servers, etc) from capistrano.yml.

rails_config:set_variables reads from capistrano.yml and sets variables and rails_config:set_servers sets the servers that multistage will use to deploy to but you'll only need to worry about rails_config:see which prints out the results of the first two rails_config tasks.

This one is very important for NYULibraries!

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/rails_config'

RVM

Loads up rvm's capistrano and hooks it on to run before deploy.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/rvm'

Send Diff

Very new idea, trys to find the difference from the last two git tags and sends the difference to a specified recipient. Run using cap send_diff:send_diff.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/send_diff'

Tagging

Tags the deployed app with a prespecified tagging convention. cap tagging:checkout_branch will specify the branch to tag and check it out. cap tagging:deploy will create the tag and deploy it.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano/tagging'

Server

Since this could work on an app that requires a server, we have some server specific capistrano tasks.

Passenger

Capistrano recipes to control passenger.

• cap deploy:start - Start passenger server.

• cap deploy:stop - Stop passenger server.

• cap deploy:restart - Restart passenger server.

• cap deploy:passenger_symlink - Links passenger to the latest app location.

  In your config/deploy.rb

  require 'nyulibraries/deploy/capistrano/server/passenger'

Puma

Capistrano recipes to control puma. Uses this very gem's rake tasks

• cap deploy:start - Start puma server.

• cap deploy:stop - Stop puma server.

• cap deploy:restart - Restart puma server.

  In your config/deploy.rb

  require 'nyulibraries/deploy/capistrano/server/puma'

Defaults

We've taken the liberty to define a default recipe and default settings. Most of our apps won't need anything past defaults, but if there are small changes you can easily override them.

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano'

This will load the default attributes, with a default recipe that will use all of the above mentioned capistrano recipes with a passenger server. All you have to do is

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano'
set :app_title, "APP_TITLE"

If your app's title is different from the git repo, then you have to also do

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano'
set :app_title, "APP_TITLE"
set :repository, "git@github.com:GITUSER/REPO_NAME.git"

Optionally, you can define who to send the git diff info to by:

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano'
set :app_title, "APP_TITLE"
set :repository, "git@github.com:GITUSER/REPO_NAME.git"
set :recipient, "email" # Has to be email address!

Finally, you can override any one of the defaults, for example overriding branch:

# In your config/deploy.rb
require 'nyulibraries/deploy/capistrano'
set :app_title, "APP_TITLE"
set :repository, "git@github.com:GITUSER/REPO_NAME.git"
set :recipient, "email" # Has to be email address!
set :branch, "cool_branch_doritos"