Rails Against The Machine

Deploying rails on Ubuntu (Dapper drake).

Well I can put it off no longer – I have to enter the “sandal and star trek T-shirt wearing” world of Linux and deployment. Soon I will be posting that Microsoft suck on Slashdot and making sniggering at newbies and calling myself 31337 Haxor. The fun filled days of actualy coding but a distant memory.

Part 1 – install a whole load of stuff

We are going to install a whole load of stuff. The default repository for Ubuntu lacks many of these so we need to enable the Universe repository.

So type: sudo gedit /etc/apt/sources.list

then uncomment the lines:

#deb http://us.archive.ubuntu.com/ubuntu dapper universe
#deb-src http://us.archive.ubuntu.com/ubuntu dapper universe

and save then update the package repository info and update all the existing packages. (this may take some time!)

sudo apt-get update
sudo apt-get dist-upgrade

So you can do everything remotely you should install SSH so you can remote in from the comfort of home (Capistrano needs it to work anyway)

sudo apt-get install openssh-server

now having done this we can install ruby and MySQL.

sudo apt-get install ruby ri rdoc mysql-server libmysql-ruby

You need to get Gem from the source forge and untar repository

sudo wget http://rubyforge.org/frs/download.php/20989/rubygems-0.9.4.tgz

tar -xvzf rubygems-0.9.4.tgz

cd rubygems-0.9.4

sudo ruby setup.rb

So we can now install rails with:  sudo gem install rails --include-dependencies

(or so we think ) ..but woe is us we get the following error message:

Bulk updating Gem source index for: http://gems.rubyforge.org

ERROR:  While executing gem ... (Gem::GemNotFoundException)

Could not find rails (> 0) in any repository

So we need to clear the gem cache so it looks in the right place:

First find the path with
gem env
then type the following to delete the cache and update it.
rm -f /path/to/gem/cache/source_cache
gem update

Now we try: sudo gem install rails --include-dependencies

and this time we succeed type the following to create a rails app by way of celebration.

rails hooray

Part 2 – install (or rather build) Mongrel

First we get a few packages required for building & compiling then we install Mongrel as a Ruby Gem

sudo apt-get install build-essential ruby1.8-dev ruby1.8

sudo gem install mongrel

test if it installed by going to the root of your rails app typing

mongrel_rails start –d

Checking if it is working by seeing if your rails application appears at http://localhost:3000  then stopping mongrel with

mongrel_rails stop

since this is a real deployment we will want a pack of mongrels so we install mongrel cluster to manage the instances

sudo gem install mongrel_cluster

Part 3 capistrano ‘fun’ .. and I mean fun in an ironic sence

First install Capistrano on your development machine

gem install Capistrano

Then ‘capify’ your application by typing:

cap --apply-to /path/to/my/app MyApplicationName

Create a ‘spinner script in the script directory.

Add some content e.g.



/Library/Rails/MYAPP/current/script/process/spawner -p 11000 -i 3

Which basically says start 3 mongrel instances on ports 11000 -11002

You then create a deploy script in deploy.rb e.g.

#to generate this file just run "capify ." from the app's root directory

#Get the list of commands with 'cap -T' 

require 'mongrel_cluster/recipes'

set :application, "myapp"

set :repository,  "svn://your.repository.com"

set :deploy_to, "/Library/Rails/#{application}"

set :mongrel_conf, "#{current_path}/config/mongrel_cluster.yml"

set :svn_username, Proc.new { "username --password svnpassword" }

#needed to avoid “no passwd entry for app! Error when starting mongrel”

#also  ailed to open `nohup.out' permission denied errors

#essentially this just determines the user who sudo is

set :runner, "root" 

# If you aren't deploying to /u/apps/#{application} on the target

# servers (which is the default), you can specify the actual location

# via the :deploy_to variable:

# set :deploy_to, "/var/www/#{application}"

# If you aren't using Subversion to manage your source code, specify

# your SCM below:

# set :scm, :subversion

role :app, "ip.of.application.server "

role :web, "ip.of.web.server"

role :db,  "ip.of.db.server" :primary => true

#This bit is required due to spinner script woes 

#since you need to grant execute access to let the spinner script run

desc "Tasks to execute after code update"

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

  run "chmod +x #{release_path}/script/spin"

  run "chmod +x #{release_path}/script/process/reaper"

  run "chmod +x #{release_path}/script/process/spawner"

  run "chmod 755 #{release_path}/public/dispatch.*"

end

To use capistrano you need to install subversion command line client and openSSH on your servers



sudo apt-get install subversion

sudo apt-get install openssh-server (if not already done)



you will also need to install on your development machine (if it is a windows one try gem install net-ssh tortoise does not count as a command line client)

Now create a directory for Capistrano to deploy to on your application server and give it write permissions so you don’t get the following error when yo do cap deploy:setup : “cannot create directory '/Library/Rails'”

sudo mkdir -p /Library

sudo chmod 777 -R /Library

Now lets create the production database on the db server with mysql and give access rights.

# On your database server

$ mysql -u root -p

Enter password: ********

Welcome to the MySQL monitor. Commands end with ; or \g.

Your MySQL connection id is 2080 to server version: 5.0.19

Type 'help;' or '\h' for help. Type '\c' to clear the buffer.

mysql> CREATE DATABASE myapplication;

Query OK, 1 row affected (0.00 sec)

mysql> GRANT ALL PRIVILEGES ON myapplication.* TO 'myapplication'@'localhost' IDENTIFIED BY 'some_pass' WITH GRANT OPTION;

Query OK, 0 rows affected (0.00 sec)

So now go to the database.yml file and enter the username and password: (best practice is to manually add in the database.yml file rather than checking into version control)

production:

  adapter: mysql

  database: your_app_production

  username: your_app

  password: password

  host: localhost

So the suggestion is to add something like this to your deploy.rb script.

task :after_update_code, :roles => :app do

db_config = "#{shared_path}/config/database.yml.production"

run "cp #{db_config} #{release_path}/config/database.yml"

end

Ok so we think we can finally deploy from our development machine with 

cap deploy:setup

but we get the following error:

connection failed for: 172.xx.x.x (Net::SSH::AuthenticationFailed: myusername)

it seems that we are trying to ssh with the same name as my development machine user. So we add the following to the deploy script

set :user, 'remote-host-username'

if we continue to have problems we could try adding some more options as recommended by the ruby wiki e.g.

 ssh_options[:port] = 22

 ssh_options[:verbose] = :debug

 ssh_options[:username] = 'remote-host-username'

 ssh_options[:host_key] = 'ssh-dss'

Success! - If you go to your application server you will see that a directory structure has been created.

Now for the next stage:

Cap deploy:cold

This time it hangs on 

** [out] Authentication realm:  some_repository

 ** [out] Username:

So cap is assuming my ssh and svn usernames are the same. 

I haven’t figured out how to deal with this – it seemed easier to create a new  matching username and get it to work that way.

Ok if that worked you can go to your mongrel instances on 11000 ect and see your application. All that remains is to link up apache as a load balancer

Part 4 cast a magic spell

Walk round your computer three times backwards reciting the following incarnation

Linux linus god of the geek

God of confusion hear me speak

sudo sudo user bugs be gone 

Bless my reference let nothing go wrong

If that didn’t work your computer is probably next to a wall – move it to the center of the room

Part 5 apache proxy

You might think that all you need to do is

sudo apt-get install apache2

then check it is running on http://localhost

…but you would be wrong.  Since we want to use it as a load balancer and proxy to the mongrel instances we need to use the proxy module but with Apache2 you need to build the source to include the proxy module !!

"Invalid command 'proxy', perhaps mis-spelled or defined by a module not included in the server configuration"

So instead we use the following commands to build from source

wget http://apache.rmplc.co.uk/httpd/httpd-2.2.6.tar.gz

tar -xvf httpd-2.2.6.tar.gz

cd httpd-2.2.6

./configure --prefix=/usr/local/apache2 --enable-mods-shared=all --enable-deflate --enable-proxy --enable-proxy-balancer --enable-proxy-http

The key parameter here being “--enable-mods-shared=all” (see http://davidwinter.me.uk/articles/2006/10/17/building-apache-22-from-source-for-ubuntu-dapper/)

But alas it still doesn’t work,  we get the error “No recognized SSL/TLS toolkit detected” so we need to install an SSL toolkit (whatever that is)

sudo apt-get install openssl libssl-dev

try again

./configure --prefix=/usr/local/apache2 --enable-mods-shared=all --enable-deflate --enable-proxy --enable-proxy-balancer --enable-proxy-http

Make

sudo make install

Now edit the configuration file along the lines of http://blog.innerewut.de/2006/04/21/scaling-rails-with-apache-2-2-mod_proxy_balancer-and-mongrel

Sudo gedit /usr/local/apache/config/httpd.conf

so apache is now installed but when we try to start it with

/usr/local/apache/bin/httpd -k start


8)The dreaded Expected x.rb to define X error

This
turned out to be because I was using a plugin for this model to
validate emails all the require statements need to be in the init file
for the plug in e.g.

require 'RFC822'
require 'common_validations'

10) rails console doesnt work
sudo ln -s /usr/bin/irb1.8 /usr/bin/irb

posted by ~J  # 10:55