This article will be updated on regular bases.

Pry

http://pryrepl.org/

Pry is awesome console tool. Onece installed you can run rails c and you can inspect objects (e.g.: ls User.last)

But key benefit is using pry as a debugger tool

# app/model/user.rb

class User < ActiveRecord::Base
  def do_stuff
    a = 3
    require 'pry'; binding.pry
    return a
  end
end

Next time you do something with User#do_stuff in web-browser, pry console will pop-up in your server (rails s) and you can debug code like with ruby debugger

ruby debugger had some issues in the past specially with maintainability. That’s why you see Ruby developers prefering pry

Edit locally used gem and rollback to original state.

Imagine you need to debug why your application fails on a gem level. If you are using bundler you can open locally used gem (Development env) like this:

EDITOR=editor_executable bundle open gem_name

Example:

cd ./my-ruby-project
EDITOR=vim bundle open sidekiq

Let say you want to know what is the value of certain argument passed during initialization of your gem. Stick a binding.pry there:

         # /home/user/.rvm/gems/ruby-2.3.0@my-ruby-project/gems/sidekiq-4.1.1/lib/sidekiq/cli.rb
@ line 364 Sidekiq::CLI#parse_config:
    358: def parse_config(cfile)
    359:   opts = {}
    360:   if File.exist?(cfile)
    361:     opts = YAML.load(ERB.new(IO.read(cfile)).result) || opts
    362:     opts = opts.merge(opts.delete(environment) || {})
    363:     parse_queues(opts, opts.delete(:queues) || [])
 => 364:     require 'pry'; binding.pry
    365:   else

Restart the server (in this case Sidekiq) and debug what you need.

Then when finished you can either manually remove your changes or run gem pristine gem_name to roll back the original state.

Example

gem pristine sidekiq
# Restoring gems to pristine condition...
# Restored sidekiq-4.1.1

Get the Active Record (Rails) db configuration

from rails c

ActiveRecord::Base.connection.instance_variable_get(:@config)

Establish connection with ActiveRecord

Sometimes you want to test the connection to your relational database (MySQL, PostgreSQL,…) without loading entire Ruby on Rails with rails c. What you can do is:

check from irb

cd ./my-rails-app && irb

require 'active_record'
require 'yaml'
require 'erb'

rails_env = 'production' # change me

conf = YAML.load(ERB.new(File.read("config/database.yml")).result)
ActiveRecord::Base.establish_connection(conf.fetch(rails_env))
ActiveRecord::Base.connection

One liner with ruby -e

# cd ./my-rails-app

bundle exec ruby -e 'require "active_record"; require "yaml"; require "erb"; ActiveRecord::Base.establish_connection(YAML.load(ERB.new(File.read("config/database.yml")).result).fetch("production")).tap { |ar| ar.connection && puts("success") }'

Other sources:

Establish connection with Redis

Sometimes you want to test the connection to your Redis serer without loading entire Ruby on Rails with rails c. What you can do is:

check from irb

cd ./my-rails-app && irb

require 'redis'

Redis.new(host: 'localhost', port:'6379', db: 0).keys[0]

One liner with ruby -e

# cd ./my-rails-app

bundle exec ruby -e 'require "redis"; Redis.new(host: "localhost", port:"6379", db: 0).keys[0]'

curl Puma/Unicorn socket

Usually your webserver (Puma, Unicorn, …) is behind loadbalancer/proxy-server like NginX.

Therefore you can do:

curl http://my-website.co

…and you NginX will forward your request to WebServer (Puma) and that will retur response again via NginX.

The problem is that simetimes you want to check if there is any problem between WebServer and NginX.

If your NginX setup is forwarding requests to localhost:3000 then no problem, all you have to do is ssh to server and do make a request from there:

ssh [email protected]
curl localhost:3000/welcome

However Common practice (I would argue best-practice) in NginX world is to forward requests to the WebServer via a socket rather than url.

Therefore you need to curl on a socket.

ssh [email protected]
curl --unix-socket /var/sockets/my-puma.sock  http:/welcome

Check your NginX configuration file to ensure where to curl to.

NginX configuration fiele is usually located in /etc/nginx/nginx.conf, /opt/nginx/nginx.conf or in sub conf. files etc/nginx/sites-enabled/default

Curl --unix-socket is there from version 7.40 so if you getting error: curl: option --unix-sock: is unknown there is another option:

# using socat
echo -e "GET /welcome HTTP/1.1\r\n" | socat unix-connect:/var/sockets/pobblecom.sock STDIO

# using netcat
echo -e "GET /welcome HTTP/1.0\r\n" | nc -U /var/sockets/my-puma.sock