DZone Snippets: capistrano code

Push your public key to a server

humblefrog (Steve) — Mon, 07 Jul 2008 11:42:26 GMT

If you've already created your ssh keys locally do this to push the public key to a server so you won't have to login to the server everytime you ssh or cap deploy.



 cat .ssh/id_rsa.pub | ssh deploy@myserver.com "cat >> .ssh/authorized_keys2"

Deploying on a passenger / mod_rails host with capistrano

jerome () — Tue, 06 May 2008 05:25:13 GMT

// add this to deploy.rb



namespace :mod_rails do

  desc <<-DESC

  Restart the application altering tmp/restart.txt for mod_rails.

  DESC

  task :restart, :roles => :app do

    run "touch  #{release_path}/tmp/restart.txt"

  end

end



namespace :deploy do

  %w(start restart).each { |name| task name, :roles => :app do mod_rails.restart end }

end

Rails Environment Specific Capistrano Includes/Overrides

justinwr (Justin R) — Sat, 05 Jan 2008 19:21:40 GMT

This cap technique allows you to include environment specific tasks and namespaces as well as override anything in the master deploy.rb. Its rather simplistic.

Add your environment specific cap scripts into RAILS_ROOT/config/deployments/.

Example command-line usage:

# RAILS_ENV=production cap deploy

...or...

# cap deploy rails_env=beta



# Example needed variables at the top of your deploy.rb



set :default_env,  'beta'

set :rails_env,     ENV['rails_env'] || ENV['RAILS_ENV'] || default_env

set :extra_deploys, 'config/deployments/'



...



# Now add this to the bottom of your deploy.rb, last thing to load.



if extra_deploys && File.exists?(extra_deploys+rails_env+".rb")

  puts "Loaded #{extra_deploys+rails_env}.rb" if load extra_deploys+rails_env

else

  puts "Could not find #{extra_deploys+rails_env}.rb"

end

You may change your default_env or any other piece to fit into your deployments. The secret is "load" method call burried in that latter code portion above. The reason you load at the end of your deploy.rb is so you can override anything defined above that call in your environment specific scripts.

Upgrade rubygems and/or specific gems themselves via capistrano

nicwilliams (Dr Nic Williams) — Wed, 19 Dec 2007 22:10:32 GMT

When a latest RubyGems is released (e.g. 0.9.5 recently) or Rails (e.g. 2.0.2) you might want to push those upgrades to all your production machines.

Add these two tasks to your deploy.rb capistrano file (add namespaces for cap2.0 if you like)



  desc "Updates RubyGems version"

  task :gem_update_system do

    sudo "gem update --system"

  end



    

  desc "Install a RubyGem from remote source"

  task :gem_install do

    puts "USAGE: GEM=gemname cap gems_install" and next unless ENV['GEM']

    sudo "gem install #{ENV['GEM']} --no-ri --no-rdoc"

  end

To upgrade rubygems and rails, for example:



cap gem_update_system

GEM=rails cap gem_install

Mongrel and Apache fun with Capistrano 2.0

hmans (Hendrik Mans) — Fri, 07 Sep 2007 16:31:34 GMT

I got the Mongrel recipes from somewhere else -- I sadly don't remember where -- and modified them a bit.



namespace :deploy do

  namespace :mongrel do

    [ :stop, :start, :restart ].each do |t|

      desc "#{t.to_s.capitalize} the mongrel appserver"

      task t, :roles => :app do

        run "mongrel_rails cluster::#{t.to_s} --clean -C #{mongrel_conf}"

      end

    end

  end

  

  namespace :apache do

    desc "Start Apache"

    task :start, :roles => :web do

      sudo "/etc/init.d/httpd start > /dev/null"

    end



    desc "Stop Apache"

    task :stop, :roles => :web do

      sudo "/etc/init.d/httpd stop > /dev/null"

    end



    desc "Restart Apache"

    task :restart, :roles => :web do

      sudo "/etc/init.d/httpd restart > /dev/null"

    end

  end



  desc "Custom restart task for mongrel cluster"

  task :restart do

    deploy.mongrel.restart

    deploy.apache.restart

  end



  desc "Custom start task for mongrel cluster"

  task :start, :roles => :app do

    deploy.mongrel.start

    deploy.apache.start

  end



  desc "Custom stop task for mongrel cluster"

  task :stop, :roles => :app do

    deploy.apache.stop

    deploy.mongrel.stop

  end



end

Dump postgres production data into development database

nicwilliams (Dr Nic Williams) — Sun, 06 May 2007 09:00:17 GMT

When bug requests come in, its great to grab a copy of the current production (or system test env) data and sync it into your development database.

Here's a capistrano recipe for postgresql:



desc "Dumps target database into development db"

task :sync_db do

  env   = ENV['RAILS_ENV'] || ENV['DB'] || 'production'

  file  = "#{application}.sql.bz2"

  remote_file = "#{shared}/log/#{file}"

  run "pg_dump --clean --no-owner --no-privileges -U#{db_user} -h#{db_host} #{db_name}_#{env} | bzip2 > #{file}" do |ch, stream, out|

    ch.send_data "#{db_password}\n" if out =~ /^Password:/

    puts out

  end

  puts rsync = "rsync #{user}@#{domain}:#{file} tmp"

  `#{rsync}`

  puts depackage = "bzcat tmp/#{file} | psql #{local_db_dev}"

  `#{depackage}`

end

Capistrano logs

nicwilliams (Dr Nic Williams) — Tue, 10 Apr 2007 20:09:05 GMT

To get the last n lines from a remote log file, use this task

Usage: cap log [-s lines=100] [-s rails_env=production]
So, "cap log" returns 100 lines from production.



desc "Returns last lines of log file. Usage: cap log [-s lines=100] [-s rails_env=production]"

task :log do

  lines     = configuration.variables[:lines] || 100

  rails_env = configuration.variables[:rails_env] || 'production'

  run "tail -n #{lines} #{app_dir}/log/#{rails_env}.log" do |ch, stream, out|

    puts out

  end

end

Capistrano : apply local patches when deploying from an external source code repository

jnewland (Jesse Newland) — Wed, 13 Dec 2006 20:51:35 GMT

I've submitted patches to a couple rails apps, and want to run off of their SCM's trunk code, but with my local patches applied. These Capistrano tasks will take any files matching

patches/*.diff

in your local directory, and apply them before restarting your app.



task :after_setup do

  patches_setup

end



task :after_update_code do

  send_and_apply_patches

end



task :patches_setup do

  run "mkdir -p #{deploy_to}/#{shared_dir}/patches" 

end



task :send_and_apply_patches do

  Dir[File.join(File.dirname(__FILE__), '../patches/*.diff')].sort.each do |patch|

    puts "sending #{File.basename(patch)}"

    put(File.read(patch),

       "#{deploy_to}/#{shared_dir}/patches/#{File.basename(patch)}",

       :mode => 0777)

    puts "applying #{File.basename(patch)}"

    run "cd #{release_path}; patch -p0 < #{deploy_to}/#{shared_dir}/patches/#{File.basename(patch)}"

  end

end

Deploying to different locations with Capistrano

bmarini (Ben Marini) — Wed, 06 Dec 2006 06:34:08 GMT

// description of your code here
In your deploy.rb file:



if ENV["SERVER"] && ENV["SERVER"] == "production"

  set :primary_server, "production.com"

  set :user, "ben"

elsif ENV["SERVER"] && ENV["SERVER"] == "staging"

  set :primary_server, "staging.local"

  set :user, "ben"

else

 ...

end



role :web, primary_server

role :app, primary_server

role :db,  primary_server, :primary => true

The you say:
rake deploy SERVER=staging

Capistrano Task Library for Deploying Subversion Tags

sava.chankov (Sava Chankov) — Mon, 23 Oct 2006 15:27:21 GMT

=begin

Capistrano Task Library > capistrano_svn_tags.rb
Created by Anil Bawa
Enhanced by Sava Chankov
Distributed Under the BSD license

+ Task library to provide subversion tag support to Capistrano.

- place this file somewhere in your rails app (i suggest lib/tasks)
- add a 'require' statement in config/deploy.rb to load this file in
- run 'rake remote:deploy_tag [tagname]' to deploy a tag (you can provide your own custom implementation of the task - edit it below)
- tag rollbacks will work fine.
- this library overrides a capistrano core method, so if you wish to use the default release type (latest revision, deployed as timestamp) then comment out the require.

=end



Capistrano.configuration(:must_exist).load do



  set :svn_tag_dir, 'tags' # new config var to denote the tag sub-directory in the svn repository  

  

  desc <<-DESC

  Update all servers with the provided tag of the source code. All this does

  is do a checkout of the provided svn tag (as defined by the svn module).

  DESC

  task :update_tag, :roles => [:app, :db, :web] do

  

    puts "  * deploying tag #{release}"



    on_rollback { delete release_path, :recursive => true }

    source.checkout_tag(self)

    run <<-CMD

      rm -rf #{release_path}/log #{release_path}/public/system &&

      ln -nfs #{shared_path}/log #{release_path}/log &&

      ln -nfs #{shared_path}/system #{release_path}/public/system

    CMD



  end



  desc <<-DESC

  Deploy svn tag, reset symlink and restart server.

  DESC

  task :deploy_tag, :roles => [:app, :db, :web] do

    set_release_tag



    transaction do

      update_tag    

      symlink

      compile_mo

    end

  end

  

  desc <<-DESC

  Deploy svn tag, migrate, compile .po files to .mo and reset symlink and restart server.

  Note that everything is done in a transaction

  DESC

  task :deploy_tag_with_migrations, :roles => [:app, :db, :web] do

    set_release_tag



    transaction do

      update_tag

      symlink

      migrate

      compile_mo    

    end

  end

  

  def set_release_tag

    set :release, ENV['TAG']

    unless release

      puts "  * no tag specified, assuming latest"

      set(:release) {source.latest_tag}

    end

  end

end



module Capistrano



  #override a configuration object method to return the correct release_path for the tag.

  class Configuration

  

    # Return the full path to the named release. If a release is not specified,

    # the provided tag name (passed as a cli parameter) is used.

    def tag_path(tag_name = release)

      File.join(releases_path, tag_name)

    end

    alias_method :release_path, :tag_path

    

  end

  

  # add a method to the subversion object to checkout tags

  module SCM

    class Subversion

      

      def latest_tag

        configuration.logger.debug "querying latest tag ..." unless @latest_revision

        tags_dir = configuration.repository + '/' + configuration.svn_tag_dir + '/'

        @latest_revision = svn_ls(tags_dir).split("\n").last.chomp('/')

      end

      alias_method :latest_revision, :latest_tag

      

      # checkout a tag from a pre-configured svn tag dir.

      def checkout_tag(actor)

        op = configuration[:checkout] || "co"

        username = configuration[:svn_username] ? "--username #{configuration[:svn_username]}" : ""

        command = "#{svn} #{op} #{username} -q #{configuration.repository}/#{configuration.svn_tag_dir}/#{configuration.release} #{actor.release_path} &&"        

        run_checkout(actor, command, &svn_stream_handler(actor))         

      end

      

    private

      def svn_ls(path)

        `svn ls #{path}`

      end

    end

  end

end