capistrano


Deploy a Rails app to Docker with Capistrano

These instructions follow my previous post on deploying multiple Rails apps with Passenger, Nginx, and Docker. Go read (and do) all that first.

Assumptions

As always, this guide assumes the production server is running Ubuntu 14.04 and has all the requisite software already installed (e.g.: Docker, Rails, Capistrano, etc.). Further, it is assumed that you have a system similar to the one described here, and that by following the instruction provided, you have a Rails application deployed in a Docker container. I will be setting up Capistrano for the Rails app in that container.

Set up project in local development environment

Update the previous Docker configuration files

Nginx configuration

Change the existing docker/my-app.conf to look like this:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
server {
    listen 80 default_server;
    listen [::]:80 default_server ipv6only=on;

    server_name example.com;

    # This used to be /home/app/my-app/public;
    root /home/app/my-app/current/public;

    # Passenger
    passenger_enabled on;
    passenger_user app;
    passenger_ruby /usr/bin/ruby2.2;
}

Change Dockerfile

Since Capistrano will be building the app, all those steps can be removed from the Dockerfile. It should now look like this:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
FROM phusion/passenger-ruby22:latest
MAINTAINER Some Groovy Cat "hepcat@example.com"

# Set correct environment variables.
ENV HOME /root
ENV RAILS_ENV production

# Use baseimage-docker's init process.
CMD ["/sbin/my_init"]

# Start Nginx and Passenger
EXPOSE 80
RUN rm -f /etc/service/nginx/down

# Configure Nginx
RUN rm /etc/nginx/sites-enabled/default
ADD docker/my-app.conf /etc/nginx/sites-enabled/my-app.conf
ADD docker/postgres-env.conf /etc/nginx/main.d/postgres-env.conf

# Install the app
ADD . /home/app/my-app
WORKDIR /home/app/my-app
RUN chown -R app:app /home/app/my-app 
RUN sudo -u app bundle install --deployment 
RUN sudo -u app RAILS_ENV=production rake assets:precompile

# Clean up APT when done.
RUN apt-get clean && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

Set up Capistrano

You work on your Rails app locally and you deploy to production. From your local development environment, go to your app’s root directory and run:

1
2
cd my-app
cap install

If Capistrano is installed (gem install capistrano), you will see something similar to this:

1
2
3
4
5
6
7
mkdir -p config/deploy
create config/deploy.rb
create config/deploy/staging.rb
create config/deploy/production.rb
mkdir -p lib/capistrano/tasks
create Capfile
Capified

This produces a pre-cooked config/deploy.rb file. For the app deployed in the previous post, change it to look like this:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
lock '3.4.0'

set :application, 'my-app'
set :repo_url, 'https://github/myprofile/my-app.git'

set :branch, 'master'
set :scm, :git

set :deploy_to, "/home/app/#{fetch(:application)}"

namespace :deploy do

  desc 'Install node modules'
  task :npm_install do
    on roles(:app) do
      execute "cd #{release_path} && npm install"
    end
  end

  desc 'Build Docker images'
  task :build do
    on roles(:app) do
      execute "cd #{release_path} && docker build -t #{fetch(:application)}-image ."
    end
  end

  desc 'Restart application'
  task :restart do
    on roles(:app) do
      execute "docker stop #{fetch(:application)} ; true"
      execute "docker rm #{fetch(:application)} ; true"
      execute "docker run --restart=always --name #{fetch(:application)} --expose 80 -e VIRTUAL_HOST=example.com --link postgres:postgres -d #{fetch(:application)}-image"
    end
  end

  before :updated, 'deploy:npm_install' 
  after :publishing, 'deploy:build'
  after :publishing, 'deploy:restart'
end

Then, in config/deploy/production.rb, modify as appropriate (it’s probably sufficient to tack this on to the end of the file):

1
server "example.com", user: "app", roles: %w{app web}

Commit and push your changes to your repository.

Set up project in production

Configuration

Before deploying with Capistrano, you need to do some configuration. Assuming that the app has already been cloned to the production machine these are the files that need adjusting:

  • my-app/config/database.yml
  • my-app/config/secrets.yml

The settings in here are not typically committed to the repository for security reasons. Assuming the Postgres configuration in the previous post, database.yml should look like this:

1
2
3
4
5
6
7
production:
  <<: *default
  database: my-app_production
  username: postgres
  password: secretp@ssword 
  host: <%= ENV['POSTGRES_PORT_5432_TCP_ADDR'] %>
  port: <%= ENV['POSTGRES_PORT_5432_TCP_PORT'] %>

secrets.yml needs to have a secret key set for production. From your app’s home directory, run:

1
rake secret

Copy the key it produces and set it in secrets.yml:

1
2
production:
  secret_key_base: PasteGeneratedKeyHere

Back in the local development environment…

From your app’s home directory:

1
cap production deploy

And now back to production…

If working with the app from the previous post, everything should be ready to go. If the site reports an error, however, you may need to setup the database in production. First, stop the Docker container:

1
docker stop my-app

Then, create and seed the database:

1
2
3
docker run --rm --link postgres:postgres my-app-image rake db:create
docker run --rm --link postgres:postgres my-app-image rake db:migrate
docker run --rm --link postgres:postgres my-app-image rake db:seed

And restart:

1
docker start my-app

Deploy a Hexo blog with Capistrano

Hexo has become a little flaky of late, but it’s still my goto software when I need to set up a new blog. It boasts One-Command Deployment, which would be great if I could figure out how to deploy it to anything other than GitHub or Heroku. There may be a way, but I’ve tried nothing and I’m all out of ideas. So instead I’ll deploy with Capistrano, because I want to try it with something other than Rails for a change.

Assumptions

You’re working on Ubuntu with the following installed on a remote machine on which to host a git repository and blog site:

Hit me up in the comments if I’ve missed any basic dependencies. The software immediately pertinent to this post (e.g., Hexo and Capistrano) will be installed as required.

I’m also assuming that you have a remote machine or cloud server on which to host a git repository and Hexo blog site. Your blog will be modified on a local machine and deployed to a production machine with Capistrano. As such, to make things easy, all the software named above needs to be installed locally and remotely.

Install Hexo on your local machine

Detailed instructions are found here, but this is how you do it in a nutshell:

1
npm install hexo-cli -g

npm should have been installed as part of the node installation.

Initialize a Hexo blog

This, of course, is not necessary if you already have a Hexo blog to work with. But if you don’t,

1
2
3
hexo init blog
cd blog
npm install

Set up a remote git repository

Capistrano talks to your blog’s remote repository when it comes time to deploy. See git remote repository SSH setup for help on how to set this up.

When the blank repository has been initialized on the remote machine, you will need to initialize git in your local Hexo blog directory (i.e., blog/ if you’re following from the previous step). This step is covered in the link provided and repeated here. Assuming you’re in the blog/ directory:

1
2
3
4
5
git init
git add .
git commit -m "Hello, my new Hexo blog"
git remote add origin git@example.com:/opt/git/my-hexo-blog.git # Change domain and project name as appropriate
git push origin master

If everything is set up correctly, you won’t even need to enter a password to push your first commit.

nginx

Add host:

1
2
3
sudo touch /etc/nginx/sites-available/my-hexo-blog.conf
sudo ln -s /etc/nginx/sites-available/my-hexo-blog.conf /etc/nginx/sites-enabled/my-hexo-blog.conf
sudo vim /etc/nginx/sites-available/my-hexo-blog.conf

Write the following to the file:

1
2
3
4
5
6
7
8
9
10
11
server {
    listen 80;
    server_name example.com www.example.com;
    access_log /var/log/nginx/example.access.log;
    error_log /var/log/nginx/example.error.log;

    location / {
        alias /home/deploy/example/current/public/;
        try_files $uri $uri/ /index.html;
    }
}

Restart:

1
sudo service nginx restart

Install Capistrano

1
gem install capistrano

Set up Capistrano

Just like hexo and git, Capistrano needs to be initialized in your project directory:

1
cap install

If successful, you will see something like this:

1
2
3
4
5
6
7
mkdir -p config/deploy
create config/deploy.rb
create config/deploy/staging.rb
create config/deploy/production.rb
mkdir -p lib/capistrano/tasks
create Capfile
Capified

With regard to the steps previously taken, modify the pre-cooked config/deploy.rb as appropriate. For example:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
set :application, "my-hexo-blog"
set :repo_url, "git@example.com:/opt/git/my-hexo-blog.git"

# Default deploy_to directory is /var/www/my_app_name
set :deploy_to, "/home/deploy/my-hexo-blog"

# ...

namespace :deploy do

  # 2015-4-14 https://gist.github.com/ryanray/7579912
  desc 'Install node modules'
  task :npm_install do
    on roles(:web) do
      execute "cd #{release_path} && npm install"
    end
  end

  desc 'Compile markdown'
  task :hexo_generate do
    on roles(:web) do
      execute "cd #{release_path} && hexo generate"
    end
  end

  before :updated, 'deploy:npm_install'
  after :deploy, 'deploy:hexo_generate'
  after :finishing, 'deploy:cleanup'
end

Then, in config/deploy/production.rb, modify as appropriate once again (out of the box, it should be sufficient to tack this on to the end of the file):

1
server "example.com", user: "deploy", roles: %w{web}

Note: the above assumes that my remote production server has a user named deploy and that this user can write to the /home/deploy/my-hexo-blog directory. Ultimately, it is up to you to determine which user deploys and where your blog is located on the file system.

Deploy

1
cap production deploy

That should do it. If something goes wrong,

1
cap production deploy --trace

will give more details.


Deploying the Rails Tutorial Sample App

I recently worked through Michael Hartl’s wonderful Ruby on Rails Tutorial as a refresher. The software implemented under his direction offers functionality that basically every modern website requires (e.g., user sign up, password retrieval, etc). That which follows documents the steps I took to deploy all the best parts of that tutorial in a production environment.

Get a server

Much of this post was ripped off from this article. They recommend Digital Ocean. I like cloudatcost.com for no other reason than because they’re cheap. For the purposes of this post, it doesn’t really matter as long as it’s installed with Ubuntu 14.04.

Add a user account

The templated Rails application is executed under this account:

1
2
3
sudo adduser deploy
sudo adduser deploy sudo
su deploy

Install Ruby

Some dependencies

1
2
sudo apt-get update
sudo apt-get install git-core curl zlib1g-dev build-essential libssl-dev libreadline-dev libyaml-dev libsqlite3-dev sqlite3 libxml2-dev libxslt1-dev libcurl4-openssl-dev python-software-properties libffi-dev

rbenv

1
2
3
4
5
cd
git clone git://github.com/sstephenson/rbenv.git .rbenv
echo 'export PATH="$HOME/.rbenv/bin:$PATH"' >> ~/.bashrc
echo 'eval "$(rbenv init -)"' >> ~/.bashrc
exec $SHELL

ruby-build plugin

1
2
3
git clone git://github.com/sstephenson/ruby-build.git ~/.rbenv/plugins/ruby-build
echo 'export PATH="$HOME/.rbenv/plugins/ruby-build/bin:$PATH"' >> ~/.bashrc
exec $SHELL

rbenv-gem-rehash plugins

1
git clone https://github.com/sstephenson/rbenv-gem-rehash.git ~/.rbenv/plugins/rbenv-gem-rehash

Ruby

1
2
3
rbenv install 2.2.1
rbenv global 2.2.1
ruby -v

bundler

1
2
echo "gem: --no-ri --no-rdoc" > ~/.gemrc
gem install bundler

The echo command prevents documentation for each gem being installed locally.

Install NodeJS

Since it is my intention to deploy this system to a production environment, I need to use the Asset Pipeline to prep my content for distribution across the web. All that requires node.

1
2
3
sudo add-apt-repository ppa:chris-lea/node.js
sudo apt-get update
sudo apt-get install nodejs

Install Rails

1
2
gem install rails -v 4.2.0
rails -v

Nginx and Passenger

Install Phusion’s PGP key to verify packages

1
2
gpg --keyserver keyserver.ubuntu.com --recv-keys 561F9B9CAC40B2F7
gpg --armor --export 561F9B9CAC40B2F7 | sudo apt-key add -

Add HTTPS support to APT

1
sudo apt-get install apt-transport-https

Add the passenger repository

1
2
3
4
sudo sh -c "echo 'deb https://oss-binaries.phusionpassenger.com/apt/passenger trusty main' >> /etc/apt/sources.list.d/passenger.list"
sudo chown root: /etc/apt/sources.list.d/passenger.list
sudo chmod 600 /etc/apt/sources.list.d/passenger.list
sudo apt-get update

nginx and passenger

1
sudo apt-get install nginx-full nginx-extras passenger

Configure

1
sudo vim /etc/nginx/nginx.conf

Uncomment the rbenv Phusion Passenger stuff. There should be some helpful hints in the file itself:

1
2
3
4
5
6
7
8
9
10
11
##
# Phusion Passenger
##
# Uncomment it if you installed ruby-passenger or ruby-passenger-enterprise
##

passenger_root /usr/lib/ruby/vendor_ruby/phusion_passenger/locations.ini;

passenger_ruby /home/deploy/.rbenv/shims/ruby; # If you use rbenv
# passenger_ruby /home/deploy/.rvm/wrappers/ruby-2.1.2/ruby; # If use use rvm, be sure to change the version number
# passenger_ruby /usr/bin/ruby; # If you use ruby from source

Get an SSL certificate

These instructions will produce a self-signed certificate:

1
2
sudo mkdir /etc/nginx/ssl
sudo openssl req -x509 -nodes -days 365 -newkey rsa:2048 -keyout /etc/nginx/ssl/nginx.key -out /etc/nginx/ssl/nginx.crt

Alternatively, validate with startssl.com for free. This document provides some excellent additional information.

Add nginx host

1
2
3
sudo touch /etc/nginx/sites-available/mydomain.conf
sudo ln -s /etc/nginx/sites-available/mydomain.conf /etc/nginx/sites-enabled/mydomain.conf
sudo vim /etc/nginx/sites-available/mydomain.conf

Write the following to the file:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
server {
    listen 80 default_server;
    listen [::]:80 default_server ipv6only=on;

    listen 443 ssl;

    server_name gofish.mobi;

    # SSL
    ssl_certificate /etc/nginx/ssl/nginx.crt;
    ssl_certificate_key /etc/nginx/ssl/nginx.key;

    # Error logs
    access_log /var/log/nginx/gofish.access.log;
    error_log /var/log/nginx/gofish.error.log;

    # Passenger
    passenger_enabled on;
    rails_env production;
    root /home/deploy/rails-tutorial-template/current/public;

    # redirect server error pages to the static page /50x.html
    error_page   500 502 503 504  /50x.html;
    location = /50x.html {
        root html;
    }

    # Static assets
    location ^~ /assets/ {
        gzip_static on;
        expires max;
        add_header Cache-Control public;
    }
}

Start, or restart nginx:

1
sudo service nginx restart

PostgreSQL

Install: 

1
sudo apt-get install postgresql postgresql-contrib libpq-dev

Create the deploy postgres user:

1
2
3
sudo su - postgres
createuser -U postgres -d -e -E -I -P -r -s deploy
exit

You’ll need to set the database password in config/application.yml.

Configure the environment

Before deploying with capistrano, a few files have to be in place. As the deploy user:

1
2
cd
mkdir -p rails-tutorial-template/shared/config

Get a secret key

If you have a rails project nearby, you can just type in

1
rake secret

Or, you can generate one by running irb

1
irb

and executing the following instructions:

1
2
3
require 'securerandom'
SecureRandom.hex(64)
exit

Copy the string generated by the SecureRandom.hex(64) command.

application.yml

This template uses figaro to manage all the sensitive stuff that sometimes goes into environment variables. The config/application.yml file it looks for isn’t committed to the repository, so you have to create it yourself:

1
2
cd rails-tutorial-template/shared/config
vim application.yml

Copy, paste, modify, and save the following:

1
2
3
4
5
6
7
8
9
10
11
12
# General
app_name: "rails_tutorial_template"

# Email
default_from: "noreply@gofish.mobi"
gmail_username: "noreply@gofish.mobi"
gmail_password: "secretnoreplypassword"

# Production
secret_key_base: "PasteTheSecretKeyFromThePreviousStepHere"
host: "gofish.mobi"
provider_database_password: "databasepassword"

I set up an account in Gmail to handle signup verifications and password resets.

database.yml and secrets.yml

There’s no sensitive information contained in the database.yml or secrets.yml files, so these can be copied directly from github.

1
2
wget https://raw.githubusercontent.com/RaphaelDeLaGhetto/rails-tutorial-template/master/config/database.yml
wget https://raw.githubusercontent.com/RaphaelDeLaGhetto/rails-tutorial-template/master/config/secrets.yml

Clone the template

This is meant to be completed on the development machine (not the server). It is assumed that postgresql and all the other dependencies are already installed (if not, do so as above).

1
2
3
4
5
6
7
git clone https://github.com/RaphaelDeLaGhetto/rails-tutorial-template.git
cd rails-tutorial-template
bundle install
sudo npm install
rake db:setup
rake db:seed
vim config/application.yml

Then copy, paste, and save the following in the file:

1
default_from: 'noreply@example.com'

Tests should all pass

1
rake

capistrano deployment

I’m still working on making this easier. From the project’s directory on the development machine set the following in config/deploy/production.rb

1
2
# Replace 127.0.0.1 with your server's IP address!
server 'gofish.mobi', user: 'deploy', roles: %w{web app}

Then run

1
bundle exec cap production deploy --trace

The deployment should succeed, but the site will not be accessible until the database is set up. Log in to the production server as deploy:

1
2
3
ssh deploy@gofish.mobi
cd rails-tutorial-template/current
RAILS_ENV=production rake db:setup

Now, enable the deploy user to restart passenger without providing a sudo password:

1
sudo visudo

Add this to the end of the file and save:

1
deploy ALL=(root) NOPASSWD: /usr/bin/passenger-config

Back on the local machine, the deployment should now succeed:

1
bundle exec cap production deploy --trace

If everything worked out right, then the app should be accessible at the configured domain name (gofish.mobi in my case).