tangramor (唐图)

自己写了个 shell 脚本来替换 Capistrano，放 github 上了

tangramor — Fri, 28 Feb 2014 09:36:33 +0800

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 ServerUser@192.168.1.61:~/id_rsa.pub
$ ssh ServerUser@192.168.1.61

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:

如何向 Ruby on Rails 分页插件 will_paginate 的链接插入其它参数

tangramor — Mon, 28 Oct 2013 18:38:39 +0800

（适用 3.0 版 will_paginate）

WillPaginate 是 RoR 里很方便的一个分页用的 gem，例如在控制器里面写

@posts = Post.paginate(:page => params[:page])

然后在视图里面写

<%= will_paginate @posts %>

就可以自动给 Post 对象分页显示了。

不过如果你想给一个通过搜索表单返回的搜索结果进行分页，使用 will_paginate 就不那么方便了，因为它不会在它的链接里自动附上搜索参数，这样除非你的模型里面明确定义了搜索方法（参考 https://github.com/mislav/will_paginate/wiki/Simple-search ），否则对一些复杂的搜索条件（例如你需要组合一些搜索条件）就无能为力了。

当然 will_paginate 考虑到了大家的需求，所以你可以通过覆盖它的 WillPaginate::ActionView::LinkRenderer 类里的方法来给它的分页链接添加额外参数。

（这里给出的方式参考了国外网友对 3.0 以下 will_paginate 的 hack）

首先，在 app/helpers 下创建文件 remote_link_renderer.rb：


# This class intends to add extra URL request parameters in pagination links  
# Usage:  
#   <%  
#    search_params = {:search => @search,  
#                    :page_num => @page,  
#                    :serial => @serial,  
#                    :code => @code,  
#                    :rule => @rule  
#                    }  
#   %>   
#   <%= will_paginate @codes, :renderer =>'RemoteLinkRenderer', :remote => search_params %>  
#  

class RemoteLinkRenderer < WillPaginate::ActionView::LinkRenderer  
  def prepare(collection, options, template)  
    @remote = options.delete(:remote) || {}  
    super  
  end  

  private  

    def link(text, target, attributes = {})  
      if target.is_a? Fixnum  
        attributes[:rel] = rel_value(target)  
        target = url(target)  
      end  

      # here to compose key-value pairs to one string like "&key1=value1&key2=value2"  
      string_remote = @remote.inject('') do |attrs, pair|  
        unless pair.last.nil?  
          attrs << %(&#{pair.first}=#{CGI::escapeHTML(pair.last.to_s)})  
        end  
        attrs  
      end  

      # append the extra parameters to target  
      attributes[:href] = target.to_s + string_remote  
      tag(:a, text, attributes)  
    end  

end

然后，在视图文件里，根据你自己的搜索表单修改 search_params 内容

<%  
  search_params = {:search => @search,  
                  :rule => @rule,  
                  :registered => @registered,  
                  :reservation => @reservation,  
                  :reg_uid => @reg_uid}  
%>  

<%= will_paginate @posts, :renderer =>'RemoteLinkRenderer', :remote => search_params %>

注意这里特地指定了 :renderer =>'RemoteLinkRenderer' 。这样你的分页链接里面都会附上你在 search_params 里面设定的参数，你就可以在控制器里面对这些参数进行处理了。

ActiveRecord 的 qurey 里，哈希形式的参数如何实现 LIKE 检索？

tangramor — Mon, 28 Oct 2013 18:35:16 +0800

http://guides.rubyonrails.org/v3.2.13/active_record_querying.html 看了下，貌似不支持 LIKE？

如何在支持 LIKE 检索的情况下避免 SQL 注入攻击？

同时使用 bootstrap-sass，bootstrap-rails

tangramor — Tue, 03 Sep 2013 18:36:49 +0800

我的环境是多用户环境（multiuser）RVM + Ruby 2.0.0 p247 + Rails 3.2.14，用了 https://github.com/RailsApps/rails3-bootstrap-devise-cancan 做模板来构建基础系统。这个模板使用的是 bootstrap-sass

在开发环境里面一切都很好，问题出在 assets:precompile，编译后的 js 不起作用，也就是说，bootstrap-dropdown, bootstrap-tab 这些在 production 环境下都不好使了。Google 了很多解决方案，试过都不好用。

然后我就想替换一下 bootstrap 实现试试，于是把 bootstrap-sass 换成了 anjlab-bootstrap-rails，assets 里面的 application.js 和 bootstrap_and_overrides.css.scss 都做了相应改动： application.js 是

//= require bootstrap

改成

//= require twitter/bootstrap

bootstrap_and_overrides.css.scss 是

@import "bootstrap";

改成

@import "twitter/bootstrap";

然后注释掉了 @import "bootstrap-responsive"; 因为 bootstrap-rails 里面没有这个

重新 bundle install 和 rake assets:precompile 后，启动 rails s -e production，发现 bootstrap-dropdown、bootstrap-tab 这样的 js 功能都可用了，但是原来的 bootstrap css 不对了，标题栏从横向变成了标准的 ul 列表从上到下显示，把页面内容都挡住了。检查了 bootstrap-rails 的 css，发现它的实现不是标准的 bootstrap css，在它的 navbar 样式表定义里（ https://github.com/anjlab/bootstrap-rails/blob/master/app/assets/stylesheets/twitter/bootstrap/_navbar.scss ）没有 .navbar-inner 这样的类定义。

再次尝试过几种 stackoverflow 上的方案都不行，我决定不使用 bootstrap-rails 的样式表（删掉 bootstrap_and_overrides.css.scss 里面的@import "twitter/bootstrap";），把 bootstrap-sass 的样式表定义（~/.rvm/gems/ruby-2.0.0-p247/gems/bootstrap-sass-2.3.2.2/vendor/assets/stylesheets/bootstrap/*）给拷贝到了 app/assets/stylesheets/bootstrap_sass 目录下，删除了 bootstrap.scss 和 responsive.scss，给除 _variables.scss 和 _mixins.scss 之外的每个文件顶部加上了

@import "_variables";
@import "_mixins";

重新 bundle install 和 rake assets:precompile（它会自动包含 app/assets/stylesheets/bootstrap_sass 目录内容）后，启动 rails s -e production，一切 OK！