https://github.com/tangramor/deploy_rails

参考了很多人的部署经验，例如蝉游记网站的部署 Nginx,Unicorn,Capistrano,OOB,Graceful Restart。因为对 Capistrano 的配置和部署方式颇感不爽，就自己用 bash 脚本实现了自动对集群的部署。

写了个使用指南，还没中文版……

deploy_rails

Use Shell script to update rails applications on remote servers through git, and execute bundle install, rake db:migration, asset precompile and restart unicorn...

I tried to use Capistrano to do the deployment, but met a lot of problems, such as configuration error, ruby version not match, code upgrade caused runtime error...then I decided to write my own script to do the job.

Prerequirements

You may need to follow steps below to deploy you ruby on rails application to a Ubuntu + Nginx + Unicorn environment. Here I installed RVM as multiple users mode. And I use www-data user to run my app (which is the user for Apache and Nginx on Ubuntu).

You can also refer to other tutorials to do the initial deployment, such as Lighting fast, zero-downtime deployments with git, capistrano, nginx and Unicorn, but it uses Capistrano, so if you follow it completely, you don't need my tool...

Or you can refer to How To Deploy Rails Apps Using Unicorn And Nginx on CentOS 6.5 if you are using CentOS. But if you are using different OS, you may need adjust some configurations in following steps accordingly.

0. Deploy source code to "production" servers

As root, change user www-data's shell to bash and set its home folder to /var/www

# vi /etc/passwd

www-data:x:33:33:www-data:/var/www:/bin/sh
==>
www-data:x:33:33:www-data:/var/www:/bin/bash

To checkout source code from git server, you need to create ssh key for www-data and add the public key string to your git server (github or private gitlab)

sudo su - www-data
$ ssh-keygen ###(Here don't input any password)
$ cat /var/www/.ssh/id_rsa.pub
......

On your production server, check out your application code from git server.

sudo su - www-data
$ git clone https://github.com/MYNAME/MYAPP.git

1. Change configuration of the application for "production" environment:

config/database.yml

Here I use MySql, you may change to your database type accordingly

production:
adapter: mysql2
encoding: utf8
reconnect: false
database: MyApp
pool: 5
username: root
password: password
socket: /var/run/mysqld/mysqld.sock
host: 192.168.1.61

config/environments/production.rb

# Disable Rails’s static asset server (Apache or nginx will already do this)
config.serve_static_assets = true

2. Setup RVM for multi-users

I use this mode because I want user www-data can use rvm directly

$ \curl -L https://get.rvm.io | sudo bash -s stable

3. Setup Ruby by RVM

$ rvmsudo rvm get head
$ rvmsudo -s rvm pkg install openssl
$ sudo su -

###(The following changes only for China mainland, because it is very slow to connect https://rubygems.org/)

# gem sources –remove https://rubygems.org/

# gem sources -a http://ruby.taobao.org/

# gem sources -l

*** CURRENT SOURCES ***
http://ruby.taobao.org
### ensure only ruby.taobao.org

# sed -i ‘s!ftp.ruby-lang.org/pub/ruby!ruby.taobao.org/mirrors/ruby!’$rvm_path/config/db

# exit

$ rvmsudo -s rvm install 2.0.0

Open /etc/bash.bashrc or /etc/bashrc (whichever is available) and add the following to the end of the file. Also run this command in your terminal.

export PATH=$PATH:/usr/local/rvm/gems/ruby-2.0.0-p247/bin
export RAILS_ENV=production
source /etc/profile.d/rvm.sh
rvm use 2.0.0 >/dev/null 2>/dev/null

Open /var/www/.bashrc and add following:

source /etc/profile.d/rvm.sh
rvm use 2.0.0 >/dev/null 2>/dev/null

4. Setup Unicorn as root

$ sudo su -
# gem install bundler
# gem install unicorn

or gem install –verbose –debug unicorn (with detailed information)

if got exception like:

Exception `OpenSSL::SSL::SSLError’ at /usr/local/rvm/rubies/ruby-2.0.0-p247/lib/ruby/2.0.0/openssl/buffering.rb:174 – read would block

Edit .gemrc :

–-
:ssl_verify_mode: 0
:backtrace: false
:benchmark: false
:bulk_threshold: 1000
:sources:
- http://ruby.taobao.org/
:update_sources: true
:verbose: true

5. Setup some required libs (for my app):

# apt-get install libmysqlclient-dev imagemagick libmagickwand-dev

6. Install bundlers as root

First, add following to Gemfile of the application:

gem 'unicorn'
gem 'unicorn-worker-killer'

Then execute following commands:

# bundle update rails
# bundle install

7. Configure Unicorn to be executed by www-data

$ vi config/unicorn.rb

# config/unicorn.rb
# Set environment to development unless something else is specified
env = ENV["RAILS_ENV"] || "development"

# See http://unicorn.bogomips.org/Unicorn/Configurator.html for complete
# documentation.
worker_processes 6

app_root = File.expand_path("../..", __FILE__)
working_directory app_root

# listen on both a Unix domain socket and a TCP port,
# we use a shorter backlog for quicker failover when busy
listen "/tmp/unicorn.socket", :backlog => 64
listen 4096, :tcp_nopush => false

# Preload our app for more speed
preload_app true
GC.respond_to?(:copy_on_write_friendly=) and
  GC.copy_on_write_friendly = true

# nuke workers after 300 seconds instead of 60 seconds (the default)
timeout 300

pid "#{app_root}/tmp/pids/unicorn.pid"

# Production specific settings
if env == "production"
  # Help ensure your application will always spawn in the symlinked
  # "current" directory that Capistrano sets up.
  working_directory app_root

  # feel free to point this anywhere accessible on the filesystem
  user 'www-data', 'www-data'

  stderr_path "#{app_root}/log/unicorn.stderr.log"
  stdout_path "#{app_root}/log/unicorn.stdout.log"
end

# Force the bundler gemfile environment variable to
# reference the Сapistrano "current" symlink
before_exec do |_|
  ENV["BUNDLE_GEMFILE"] = File.join(app_root, 'Gemfile')
end

before_fork do |server, worker|
  # the following is highly recomended for Rails + "preload_app true"
  # as there's no need for the master process to hold a connection
  if defined?(ActiveRecord::Base)
    ActiveRecord::Base.connection.disconnect!
  end

  # Before forking, kill the master process that belongs to the .oldbin PID.
  # This enables 0 downtime deploys.
  old_pid = app_root + '/tmp/pids/unicorn.pid.oldbin'
  if File.exists?(old_pid) && server.pid != old_pid
    begin
      Process.kill("QUIT", File.read(old_pid).to_i)
    rescue Errno::ENOENT, Errno::ESRCH
      # someone else did our job for us
    end
  end
end

after_fork do |server, worker|
  # Disable GC, together with the OOB after to reduce the execution time
  GC.disable
  # the following is *required* for Rails + "preload_app true",
  if defined?(ActiveRecord::Base)
    ActiveRecord::Base.establish_connection
  end

  # if preload_app is true, then you may also want to check and
  # restart any other shared sockets/descriptors such as Memcached,
  # and Redis.  TokyoCabinet file handles are safe to reuse
  # between any number of forked children (assuming your kernel
  # correctly implements pread()/pwrite() system calls)
end

$ vi config.ru

# This file is used by Rack-based servers to start the application.

require ‘unicorn/oob_gc’
require ‘unicorn/worker_killer’

# execute 1 GC every 10 requests
use Unicorn::OobGC, 10

# Set the max request time to avoid mem leak by GC (random from 3072 to 4096 so the processes will not kill themselves at the same time)
use Unicorn::WorkerKiller::MaxRequests, 3072, 4096

# Set the max memery size to avoid mem leak by GC (random from 192 to 256 MB so the processes will not kill themselves at the same time)
use Unicorn::WorkerKiller::Oom, (192*(1024**2)), (256*(1024**2))

require ::File.expand_path(‘../config/environment’,??__FILE__)
run MyApp::Application

8. Configure Nginx

# vi /etc/nginx/sites-enabled/default

upstream myapp {
    # fail_timeout=0 means we always retry an upstream even if it failed
    # to return a good HTTP response (in case the Unicorn master nukes a
    # single worker for timing out).
    # for UNIX domain socket setups:
    server unix:/tmp/unicorn.socket fail_timeout=0;
}

server {
    listen   80; ## listen for ipv4; this line is default and implied
    #listen   [::]:80 default_server ipv6only=on; ## listen for ipv6

    root /var/www/MYAPP/public;
    index index.html index.htm;

    # Make site accessible
    server_name myhostname.com;

    location / {
        proxy_pass  http://myapp;
        proxy_redirect     off;

        proxy_set_header   Host             $host;
        proxy_set_header   X-Real-IP        $remote_addr;
        proxy_set_header   X-Forwarded-For  $proxy_add_x_forwarded_for;

        client_max_body_size       10m;
        client_body_buffer_size    128k;

        proxy_connect_timeout      90;
        proxy_send_timeout         90;
        proxy_read_timeout         90;

        proxy_buffer_size          4k;
        proxy_buffers              4 32k;
        proxy_busy_buffers_size    64k;
        proxy_temp_file_write_size 64k;

    }

    location ~ ^/(assets)/{
        root /var/www/MYAPP/public;
        expires max;
        add_header Cache-Control public;
    }

    location ~* \.(jpg|jpeg|gif|png|ico|css|bmp|js|html)$ {
        root   /var/www/MYAPP/public;
    }
}

Deploy the script

1. Copy unicorn scritp to /etc/init.d/unicorn and make it executable

chmod +x /etc/init.d/unicorn

2. Edit /etc/init.d/unicorn to match your environment

USER="www-data"
DAEMON=unicorn
RUBY_VERSION=2.0.0
RAILS_ENV=production
RVM_PROFILE="/etc/profile.d/rvm.sh"
APPPATH="/var/www/MYAPP"
DAEMON_OPTS="-c $APPPATH/config/unicorn.rb -E $RAILS_ENV -D"
NAME=unicorn
DESC="Unicorn app for $USER"
PID=$APPPATH/tmp/pids/unicorn.pid

In most cases, you just need to modify USER, RUBY_VERSION, RAILS_ENV, RVM_PROFILE and APPPATH to the values of your server

3. Set unicorn auto start when boot up

# update-rc.d unicorn defaults

4. Configure SSH connection between workstation and servers

You need to generate ssh key on your workstation. For example, my work machine is a laptop and I installed Ubuntu 13.10 on it, so I generate ssh key by execute ssh-keygen without input passphrase for the key (this will generate a key with no passphrase)

$ ssh-keygen
Generating public/private rsa key pair.
Enter file in which to save the key (/home/USERNAME/.ssh/id_rsa):
......

Then append the public key string to the servers. In this case we use www-data user, so just log in the server, open /var/www/.ssh/authorized_keys and add the content of /home/USERNAME/.ssh/id_rsa.pub on you workstation to it:

On your workstation------------------------------------------
$ scp ~/.ssh/id_rsa.pub [email protected]:~/id_rsa.pub
$ ssh [email protected]

Now we are on 192.168.1.61-----------------------------------
$ cat ~/id_rsa.pub | sudo -u www-data sh -c "cat - >>/var/www/.ssh/authorized_keys"
$ rm ~/id_rsa.pub

5. Deploy code changes

After you committed any change on you app and pushed to your git server, you can execute update_unicorn.sh to deploy the changes to your production servers. You need to edit update_unicorn.sh to match your environment.

SERVER_IPS=(192.168.1.61 192.168.1.62)
SERVER_USERS=(www-data www-data)
RUBY_VERSION=2.0.0
RVM_PROFILE="/etc/profile.d/rvm.sh"

In this example we have 2 servers: 192.168.1.61 and 192.168.1.62, they are 2 nodes of a cluster. You can reduce or add servers here. And on both servers we use www-data user to execute the rails application.

Reference:

1. http://ihower.tw/rails3/deployment.html

2. http://ruby-china.org/topics/12033

3. http://ruby.taobao.org/

4. http://ariejan.net/2011/09/14/lighting-fast-zero-downtime-deployments-with-git-capistrano-nginx-and-unicorn/