Unixmonkey.net

For some complex searches, I find myself needing to build a query with some optional conditions that may or may not exist based on search terms.

I could concatenate an SQL string, but when we are dealing with user-supplied input, we need to parameterize the sql query with question marks (?) for binding parameters to to avoid SQL injection attacks.

But I may not know how many parameters I’m actually going to use in a query.

With a form like this:

Show me my friends:
  living in: [   ]
  whose hobby is: () skating, () fishing, () basketball
  over: [  ] years old

  *(no required fields)

At first stab, you could try something like this:

User.friends.find(:all, :conditions => [
   'town LIKE ? AND hobby = ? AND age >= ?', 
   "%#{params[:town]}%", params[:hobby], params[:age] 
])

This would work, but only if all fields are filled out; otherwise the SQL generated would break looking like this if someone only filled in the name part of the form:

"SELECT * FROM users WHERE town LIKE "%indianapolis%" AND hobby = '' AND age >= ''"

There are some plugins like Ezra’s Ez-Where to handle query building; but how about we try using the built-in symbol key interpolation as described here in ActiveRecord to get the job done.

This lets you use named :symbols in place of question marks for binding.

Find lets you supply your conditions as a 2-element array with a string, and a corresponding hash like so:

User.friends.find(:all, :conditions => [ 
  'town LIKE :town AND hobby = :hobby AND age >= :age', 
  { :town => "%#{params[:town]}%" , :hobby => params[:hobby],  :age = params[:age] }
])

Again, this will work just fine if all fields are filled out; but how do we omit conditions and hash keys?

Our final SQL string can be built by joining conditions with AND (or OR if your app suggests it), and stuffing new key/value pairs into our arguments hash, or with Hash.merge!

First we’ll set up our search method like so:

def results
  conditions  = []
  arguments = {}
 
  unless params[:town].blank?
    conditions << 'town LIKE :town'
    arguments[:town] = "%#{params[:town]}%"
  end
 
  unless params[:hobby].blank?
    conditions << 'hobby = :hobby'
    arguments[:hobby] = params[:hobby]
  end
 
  unless params[:age].blank?
    conditions << 'age >= :age'
    arguments[:age] = params[:age]
  end
 
  all_conditions = conditions.join(' AND ')
 
  @user_friends = User.friends.find(:all, :conditions => [all_conditions, arguments])
end

An alternative is to use Hash.merge! like so:
(this will let you assign more than one key/value pair at a time or to combine hashes)

instead of:

arguments[:town]  = "%#{params[:town]}%"
arguments[:hobby] = params[:hobby]

  arguments.merge!({ 
    :town  => "%#{params[:town]}%",
    :hobby => params[:hobby]
  })

Maybe not as elegant as using a plugin, but certainly clear and flexible.

I’ve been playing around with Shoes (shoooes.net) lately as a way to put a cross-platform graphical user interface (GUI) on some of my small purpose-built command-line ruby scripts.

I find that it is quite easy to get started with, and lends a lot of flexibility to the way your program is structured and displayed. However, the structure feels a little bit alien compared to everyday ruby, and there are some gotcha’s you need to keep in mind while developing for Shoes.

I feel I must preface this article by saying that Shoes has excellent documentation, _why (the lucky stiff) turns documentation into its own art form. The manual, “Nobody Knows Shoes” reads a lot like a comic book, full of _why’s own original artwork and clippings from old-timey photos and art, and is complimented by the documentation at help.shoooes.net

I had a bit of trouble at first getting ActiveRecord to interface with a database from a straight port from one of my console apps because I glossed over the parts of the manual that detail the tricky behavior of the garbage collector reaping predefined classes after the app’s initial load.

The fix is pretty simple. Stick all your classes in an external file (or many) and load them using ‘require’.

Anyhow, here is a barebones example of a working implementation for using ActiveRecord in Shoes:

# in foo.rb
class Foo < ActiveRecord::Base
end
 
#in app.rb
Shoes.setup do
  gem 'activerecord'
  require 'active_record'
  ActiveRecord::Base.establish_connection(
    :adapter   => 'sqlite3',
    :dbfile    => 'foos_db.sqlite3'
  )
  require 'foo'
end
Shoes.app do
  @foos = Foo.find(:all)
  para @foos
end

Now, this example requires there is an existing sqlite database with a foos table, change out the establish_connection parameters to connect to any other database. The gem ‘activerecord’ statment tells shoes to install the activerecord gem into the shoes ruby library if it isn’t already there.

If you don’t already have a database, and just want to use a db to act as a storage layer for your app, then you might want to use ActiveRecord::Schema.define to create a database and setup the tables the same way you do for Rails migrations.

Here is a more complete example of an app to keep track of notes using ActiveRecord as the backend. I like the “base class that inherits from Shoes” pattern, so I’m using that here.

# in note.rb
class Note < ActiveRecord::Base
end
 
# in app.rb
Shoes.setup do
  gem 'activerecord' # install AR if not found
 
  require 'active_record'
  require 'fileutils'
 
  ActiveRecord::Base.establish_connection(
    :adapter   => 'sqlite3',
    :dbfile    => 'shoes_app.sqlite3'
  )
 
  # create the db if not found
  unless File.exist?("shoes_app.sqlite3")
    ActiveRecord::Schema.define do
      create_table :notes do |t|
        t.column :message, :string
      end
    end
  end
 
end
 
class ShoesApp < Shoes
  require 'note'
 
  url '/', :index
 
  def index
    para 'Say something...'
    flow do
      @note = edit_line
      button 'OK' do
        Note.new(:message => @note.text).save
        @note.text = ''
        @result.replace get_notes  
      end
    end
    @result = para get_notes
  end
 
  def get_notes
    messages = []
    notes = Note.find(:all, :select => 'message')
    notes.each do |foo|
      messages << foo.message
    end
    out = messages.join("\n")
  end
 
end
 
Shoes.app :title => 'Notes', :width => 260, :height => 350

Here’s a screenshot:

There you are; a cross-platform desktop app that doesn’t require a full-on build environment, and can be distributed with the source exposed for later improvements.

The first time this runs, it installs Activerecord, requires it, establishes a connection, creates the table unless one already exists. Then it shows a form to add notes followed by all the existing notes in the database. Adding a new note refreshes the notes shown.

This isn’t exactly a polished app with full CRUD, but should prove a good introduction to Shoes for someone used to working with ActiveRecord.

I’ve got some data in an excel file that I need to put in the database and its far too much to do by hand, what will I do?

Lets throw some ruby at the problem!

First, excel it too darn complicated and proprietary a format to even mess with unless you are creating something really worth it, so lets open that .xls with Excel or OpenOffice and do a File -> Save As -> .csv (comma separated values) to get a file that is easier to work with.

Now, we could write our own CSV parser since its such a simple format, but why futz with it when someone else has already put out a good library for that will likely be more error tolerant? Lets use FasterCSV, as its pretty well-known.

Install by issuing:

sudo gem install fastercsv

Now you can just fire up script/console of your Rails app and type in the below, or just put this in a database migration to slurp up all that good spreadsheet data.

The below assumes you have a ‘users’ table with fields name, address, and email that are also rows in your excel file. Adjust as necessary.

require 'fastercsv'
FasterCSV.foreach("#{RAILS_ROOT}/myfile.csv") do |row|
  record = User.new(
    :name    => row[0], # first column of csv file
    :address => row[1], # second column
    :email   => row[2]  # third
  )
  record.save
end

That’s pretty awesome; now how can I export that stuff in the database back out to Excel again?

Lets use Ruport, the Ruby report gem!

sudo gem install acts_as_reportable

Toss the require statement somewhere obvious (like environment.rb or above the model you want to export), and put ‘acts_as_reportable’ in your model declaration.

require 'ruport'
class User < ActiveRecord::Base
  acts_as_reportable
end

Now I can do this kind of stuff to export to a csv file (again with script/console, but a migration should work equally well):

content = User.report_table.as(:csv) # convert your model table to CSV

or

content = User.report_table_by_sql("SELECT name, address, email FROM users").as(:csv)

Then write that to a file like so:

file = File.open("#{RAILS_ROOT}/report.csv", "w") # open file
file.print(content) # print that csv content to the open file
file.close          # close the file

Open that CSV file with Excel and amaze the sales team, your boss, or whoever.

This is a narrow view of what we can do with FasterCSV and Ruport, but I’m sure you can see how you could build out a format.csv in a respond_to block in a Rails controller, or have a setter in your model that sucks in an uploaded CSV to create some records.

These are some pretty great libraries, and I’m very glad they were able to help me load, combine, query and output some data I’d been working with in a pinch.

I hope this post serves to help someone else in a similar situation.

Sure, I haven’t even turned it on in more than 5 years, and I certainly didn’t use it very much back then either. Its been sitting in the bottom of the storage closet doing no one any good, yet it pains me just a bit to let it go.

Not because I won’t ever be able to turn it on again, as it was pretty likely I never again would, but because just having it gave me a small sense of “Hey, I was there. In that short time where computing had been pushed into the living rooms before it was practical, before the internet as we know it existed, where operating the computer was only a few steps away from programming it.”

The Commodore 128 was a gift to me from a lawyer who also hadn’t used it in years, but used it as a high-end word processor, as the backspace keys on typewriters left a tell-tale sign that you had definitely erased something (and with a pocketknife, you could probably tell what).

It wasn’t my first Commodore, though. I got a secondhand Vic-20 when I was about 8. It came with a manual that was little more than a “How to program basic” with some sample code to write some simple games, and some atari-like cartridges with games like “space invaders”, and “adventure”. I was a big fan of the text-based games. I later ended up tearing it apart to check out its innards, and left it broken for too long that circuit boards and small parts went missing.

I remember getting in trouble for “hacking” in school the first time we were taken to the new computer lab equipped with early apple’s running basic, and I had instructed the computer to run an infinite loop of the below program I had learned from my trusty Commodore manual.

10 PRINT "TOMMY STINKS"
20 GOTO 10

I had a friend whose father was an avid computer geek in these mid-to-late 80’s days, who would pirate video games by tuning his ham-radio to a station that broadcast the analog signal he could tape onto a cassette he could read back into the computer.

I don’t know if I’m reminiscing, or just trying to hold on to the memory of the way things were because they were so drastically different from today. The information economy and the internet as we know it has forever changed the way we work, play, and live.

Anyhow, I’ll miss the Commodore all the same; but I’m happy it found a new home, freed up some of my closet space, and helped me put down the deposit on my new apartment.

I took a bunch of pictures before putting it up on eBay, you can see them all here: http://unixmonkey.net/Commodore/

The web is a generally free and open place for all types of communication, but if you put your email address on 1 website, you can expect an email-harvesting robot spider to find that address and send it to its spammer overlords.

Once on a spammer’s list, you can expect to get all kinds of interesting stock tips, products to enhance your manhood, and friendly letters from Nigerian diplomats.

If you simply have too little to do in the day, this can be a great way to meet new people and start a career in day trading. However, some of us are just too darn busy to stop what we are doing every 2/3rds of a second to check our email; but still need it for keeping in contact with friends, family, and business contacts.

From a few tips pulled from the web, I set to create a nice link helper for Ruby / Rails intended to display email links that work indistinguishably from regular mailto: links, and even gracefully downgrade for users without javascript.

Lets not even display the email address on the page at all, and use a little javascript to render the email address after the fact by breaking it up and putting it back together with javascript.

# Takes in an email address and (optionally) anchor text,
# its purpose is to obfuscate email addresses so spiders and
# spammers can't harvest them.
def js_antispam_email_link(email, linktext=email)
    user, domain = email.split('@')
    # if linktext wasn't specified, throw email address builder into js document.write statement
    linktext = "'+'#{user}'+'@'+'#{domain}'+'" if linktext == email 
    out =  "<noscript>#{linktext} #{user}(at)#{domain}</noscript>\n"
    out += "<script language='javascript'>\n"
    out += "  <!--\n"
    out += "    string = '#{user}'+'@'+'#{domain}';\n"
    out += "    document.write('<a href='+'ma'+'il'+'to:'+ string +'>#{linktext}</a>'); \n"
    out += "  //-->\n"
    out += "</script>\n"
    return out
end

This is probably good enough for 90% of those robots, but you know if one spammer gets your address, he will likely share (or sell) your email to all his friends. The weak spot in this looks like the noscript version, lets fuzz that up a bit by converting to HTML character entities.

One of the earliest and simplest ways to obfuscate an email address is by converting each character into its HTML equivalent. This makes the source look nasty, but will be correctly rendered by the browser that the end-user is none the wiser.

An address like abc@example.com will look like this in the source:

abc@example.com

Let’s build a simple method to convert a plaintext string into something like the above. I’m going to cheat and only convert a-z and A-Z and leave @ signs, dots, dashes, etc. alone.

# HTML encodes ASCII chars a-z, useful for obfuscating
# an email address from spiders and spammers
def html_obfuscate(string)
  output_array = []
  lower = %w(a b c d e f g h i j k l m n o p q r s t u v w x y z)
  upper = %w(A B C D E F G H I J K L M N O P Q R S T U V W X Y Z)
  char_array = string.split('')
  char_array.each do |char|  
    output = lower.index(char) + 97 if lower.include?(char)
    output = upper.index(char) + 65 if upper.include?(char)
    if output
      output_array << "&##{output};"
    else 
      output_array << char
    end
  end
  return output_array.join
end

now in our js_antispam_email_link method we can “encrypt” the user and domain before sending to the browser like so:

def js_antispam_email_link(email, linktext=email)
  user, domain = email.split('@')
  user = html_obfuscate(user)
  domain = html_obfuscate(domain)
  ...

Not bad, but many spiders these days can still decode HTML entities and get at that address, so lets build up our defenses a bit more by adding some methods to really screw with those spiders.

We’ll write a method that encrypts a string with ROT13 and puts that on the webpage, and use some javascript to decrypt that on page display. ROT13 is a really simple cipher where you take characters a-z and shift them by half the alphabet.

This is a really simple one-liner borrowed from Jay Komineck

# Rot13 encodes a string
def rot13(string)
  string.tr "A-Za-z", "N-ZA-Mn-za-m"
end

Lets use this to really beef up our link helper by using some javascript that can decipher this. JS code taken from Allan Odgaard

string = '#{email}'.replace(/[a-zA-Z]/g, 
  function(c){ 
    return String.fromCharCode(
      (c <= 'Z' ? 90 : 122) >= (c = c.charCodeAt(0) + 13) ? c : c - 26
    );
  });

Now we’ve got some pretty strong defense against those pesky robots and by using simple HTML character encoding and lightweight ROT13 ciphering it shouldn’t be too taxing on your webserver to spit out a page with a few emails on it. Less sophisticated browsers still get the contact info and everyone is a little bit happier to come home to a (relatively) clean inbox.

Here’s the whole shebang put together, put this in application_helper.rb if using rails:

# Rot13 encodes a string
def rot13(string)
  string.tr "A-Za-z", "N-ZA-Mn-za-m"
end
 
# HTML encodes ASCII chars a-z, useful for obfuscating
# an email address from spiders and spammers
def html_obfuscate(string)
  output_array = []
  lower = %w(a b c d e f g h i j k l m n o p q r s t u v w x y z)
  upper = %w(A B C D E F G H I J K L M N O P Q R S T U V W X Y Z)
  char_array = string.split('')
  char_array.each do |char|  
    output = lower.index(char) + 97 if lower.include?(char)
    output = upper.index(char) + 65 if upper.include?(char)
    if output
      output_array << "&##{output};"
    else 
      output_array << char
    end
  end
  return output_array.join
end
 
# Takes in an email address and (optionally) anchor text,
# its purpose is to obfuscate email addresses so spiders and
# spammers can't harvest them.
def js_antispam_email_link(email, linktext=email)
  user, domain = email.split('@')
  user   = html_obfuscate(user)
  domain = html_obfuscate(domain)
  # if linktext wasn't specified, throw encoded email address builder into js document.write statement
  linktext = "'+'#{user}'+'@'+'#{domain}'+'" if linktext == email 
  rot13_encoded_email = rot13(email) # obfuscate email address as rot13
  out =  "<noscript>#{linktext}<br/><small>#{user}(at)#{domain}</small></noscript>\n" # js disabled browsers see this
  out += "<script language='javascript'>\n"
  out += "  <!--\n"
  out += "    string = '#{rot13_encoded_email}'.replace(/[a-zA-Z]/g, function(c){ return String.fromCharCode((c <= 'Z' ? 90 : 122) >= (c = c.charCodeAt(0) + 13) ? c : c - 26);});\n"
  out += "    document.write('<a href='+'ma'+'il'+'to:'+ string +'>#{linktext}</a>'); \n"
  out += "  //-->\n"
  out += "</script>\n"
  return out
end

I hope this helps out somebody out there, please leave a comment if you have any suggestions.

If you’ve ever wanted to keep track of revisions to document files or images in your Rails app, you are likely to want to use Acts_as_versioned, which is the authority on versioning database records, and Attachment_fu, which is the authority on uploading files with Rails.

The problem is that they don’t know about each other and will step on each other’s toes without some changes. This article serves as a quick introduction to each, and shows how to make the two plugins get along like best friends.

Acts_as_versioned was written by Rails Core Team member Rick Olsen (who also wrote attachment_fu and Restful_authentication among others) that essentially makes a mirror table of the one you want to version, and keeps every version of the record you are updating.

Say I have a document table with fields like this:

Acts_as_versioned will add a column “version”, and a separate table “document_versions”.

The document_versions table will look a bit like this

Setting up acts_as_versioned is pretty simple, I got most of my introduction to it from urbanhonking.com

Now every time you update the original document, the changes are saved in your main documents table, and the version column is incremented by 1.

After a few edits of the document, you’ll see the versioning information in the Document_versions table add up.

Great! We can now use some of acts_as_versioned’s built-in methods for determining if there are older versions, and be able to view or even revert to them.

Now lets add the ability to upload a file to attach to a document record with attachment_fu.

Attachment_fu is another plugin that makes uploading files and keeping track of them in the database relatively simple.

A good intro to attachment_fu can be found on Mike Clark’s blog

Attachment_fu would require a few changes to our documents table:

Don’t forget to add the same fields to your documents_versions table, too.

Once we’ve added the right file fields to the new and edit forms, and image_tag or download link on the show view, we’ve got working file uploads. Nice.

Try to edit a record by attaching a new file, the new file is displayed and the record is preserved as an older version in the versioned table. But if you try to view the old version…wait a minute? Where did my version 1 file go!

That’s right, attachment_fu deletes the old file when you add a new one (as it should if you aren’t versioning your data). Attachment_fu’s rename_file method is the one responsible for deleting (or renaming) the old file when a new one is added, so lets monkeypatch that in our model to not do anything.

def rename_file
end

Now, it will only overwrite the file if the filename is the same. Lets store each version in its own folder to keep them from clobbering each other by monkey-patching the path files get written to in our model also:

def attachment_path_id
  "/#{id}/v#{version}/"
end
def partitioned_path(*args)
  attachment_path_id + args.to_s
end

This changes the public path from /0000/0001/rep08.jpg to /1/v1/rep08.jpg

Now, if we want to display the image, we cannot use the ‘public_filename’ method, because it is only given to the Document model, and not the Document_Version model.

That’s okay, because with our new path arrangement, we can reliably predict where the old versions of the files will be kept. You can show them with some code similar to this in your views:

<% for version in @document.versions %>
 
  Version <%= version.version %>
  <%= image_tag("/documents/#{@document.id}/v#{version.version/" + version.filename) %>
  <hr />
 
<% end %>

Now, when we delete a record, attachment_fu only knows about the current document, and will leave behind orphaned files and folders from the old versions. Lets fix that by having it get rid of the document id folder.

Rails reserves some special methods (callbacks) for performing actions before or after other major actions, lets tap into that by defining a method that will magically get called every time we delete a record.

def after_destroy
  FileUtils.rm_rf(RAILS_ROOT + "/public/documents/#{id}/")
end

This translates into the shell command rm -rf and deletes our ID directory and everything inside it.

Hooray!

As a wrap up, lets look at our complete Document model:

class Document < ActiveRecord::Base
  acts_as_versioned
  has_attachment :storage => :file_system
 
  def rename_file
  end
 
  def attachment_path_id
    "/#{id}/v#{version}/"
  end
 
  def partitioned_path(*args)
    attachment_path_id + args.to_s
  end
 
  def after_destroy
    FileUtils.rm_rf(RAILS_ROOT + "/public/documents/#{id}/") if id
  end
 
end

I’ve whipped up a sample Rails app demonstrating the points and code in this article. It uses Rails 2.0.2 with the sqlite3 database.

Download it here: Attachments_versioned (240kb .zip)

I hope this saves some work for someone who wants to leverage these two excellent plugins by Rick Olsen (technoweenie) on the same model without having them fight too much.

Autotest is part of the ZenTest suite for testing ruby and rails apps. Instead of running ‘rake test’ every time you want to run your tests, simply open another tab in your terminal, browse to your rails project directory and run ‘autotest’. It will run your test suite and sit there waiting for any file to be modified. When you edit any file in your rails project, autotest will automatically run your tests again.

To install ZenTest and autotest, open a terminal and run ’sudo gem install ZenTest’.

This is great, awesome even. I don’t want to understate how useful this is, its like breathing when doing test-driven development, but when I’m coding, I’ll often have a different terminal up front with a script/console shell or tail-ing the development log. I want my tests to get in my face when something goes wrong.

The way to do this is with Growl on OSX, Snarl for Windows, or several other similar pop-up notification apps.

As of this writing, the latest version of Growl is 1.1.2, the ZenTest Gem is 3.8.0, and OSX Leopard is 10.5.1. I installed Leopard as a clean install.

After downloading and installing Growl, while the .dmg is still mounted, open a terminal and run ‘/Volumes/Growl\ 1.1.2/Extras/growlnotify/install.sh’.

This will put a commandline tool called ‘growlnotify’ in your /usr/local/bin dir that is used to pop up those messages. Now you can use a text editor to create a ~/.autotest file.

Lets start simple and check that growl is working with autotest, add this line to your .autotest file:

require ‘autotest/growl’

Then save, and go into your Rails project and run ‘autotest’

If all the pieces were installed right, you should get some nice popups when you run autotest and each time you modify a file in your rails project directory.

Lets pretty this up a bit by over-riding the Autotest::Growl module in our .autotest file. I’ve pretty much copied this verbatim from /Library/Ruby/Gems/1.8/gems/ZenTest-3.8.0/lib/autotest/growl.rb
only I’ve added an image directive to the autotest hooks, a nice red rails logo for test failures, and a green one for test passes, then changed the growl method to substitute my images into the growlnotify command. Be sure the modify the image paths for your images directory.

# ~.autotest
 
require 'autotest/redgreen'
require 'autotest/growl'
 
module Autotest::Growl
 
  def self.growl title, msg, img="/Applications/Mail.app/Contents/Resources/Caution.tiff", pri=0
    title += " in #{Dir.pwd}"
    msg += " at #{Time.now.strftime("%I:%M %p")}"
    system "growlnotify -n autotest --image #{img} -p #{pri} -m #{msg.inspect} #{title}"
  end
 
  Autotest.add_hook :run do  |at|
     growl "autotest running", "Started"
   end
 
   Autotest.add_hook :red do |at|
     img = "/Users/djones/.autotest_images/rails_fail.png"
     growl "Tests Failed", "#{at.files_to_test.size} tests failed", img, 2
   end
 
   Autotest.add_hook :green do |at|
     img = "/Users/djones/.autotest_images/rails_ok.png"
     growl "Tests Passed", "Tests passed", img, -2 if at.tainted
   end
 
   Autotest.add_hook :all_good do |at|
     img = "/Users/djones/.autotest_images/rails_fail.png"
     growl "Tests Passed", "All tests passed", img, -2 if at.tainted
   end
 
end

Now create a .autotest_images folder in your home directory and put these images in there (I got them from here), or you can use your own.

We all know mongrel is the bees knees when it comes to serving Rails or Merb apps, even if its creator had a meltdown.

But there is a new kid on the block named thin that claims to run Rails apps almost twice as fast as mongrel. Check the graph.

A simple ‘gem install thin’ and going into your rails app and issuing ‘thin start’ is enough to get you shedding weight.

This is definitely something I’ll be playing with for the next few days.

Pushing your app to the real world with Rails can be a very daunting task to someone who is used to just uploading flat html or php files.

If you’ve never dealt with Apache, mod_rewrite, or proxy servers, prepare to spend a lot of time hammering out a solution. Worried about performance, or your need to scale out at a later date? Pick a solution that won’t leave you wondering if your site will be able to take a spike in traffic.

For Rails, there are lots of deployment strategies, some are tuned for compatibility with shared hosting (fastcgi), and some are built for speed and minimum configuration (Litespeed Nginx and Lighttpd), and some are built to leverage the strength and flexibility of stable and established server software (Apache+Mongrel).

If you want to scale, you’re going to need a load balancing proxy. The choices here are pound, pen, and apache’s mod_proxy_balancer.

An abundance of choice is a double-edged sword. Competition is good in any arena, but it makes it very hard for someone who hasn’t tried them all to choose one.

I’ve chosen to setup my Rails server using the Apache+Mongrel+mod_proxy_balancer combination.

For the operating system, I’ve chosen the newest Ubuntu 7.10 (Gusty) server. Ubuntu uses the Debian-style apt package management, but with more current packages than Debian stable, and is the current cool kid on the block for Linux systems. I have been using Ubuntu for years and can attest to its stability and cutting edge (but not bleeding edge) packages.

I’ve gone through and built a Rails server by hand before, but poorly documenting what I had done, and not securing it properly. When I stumbled upon Slicehost’s server setup articles, I knew I had found exactly what I needed to build a server configured like a pro, and all the documentation I should have written in the first place.

Here is the step-by-step setup for a production Rails server. I’ve tested this setup and can vouch for its awesomeness. I encourage you to make some changes specific to your setup where appropriate. You should be able to skip unneccesary stuff like php and virtual hosts if you don’t need them.

Ubunty Setup - part 1

Ubunty Setup - part 2

Myql and Rails

Apache and PHP

Apache Virtual Hosts

Vhosts and permissions

Apache Vhosts, Rails and Mongrels

Mongrel Clustering

Subversion intro

Setting up svnserve

Securing svnserve with ssh

Setting up Capistrano

Setting up Piston to manage plugins

As you can see, most of these are from Slicehost’s documentation articles. They also detail setup for several other major linux distributions and other deployment strategies like nginx. Major props to slicehost for putting such excellent documentation together.

Twas the morning before Christmas, and all through the office, not a creature was stirring, but me and my mouse. The code was checked in to the repo with care, in hopes a new feature soon would be there.

I sit in my sweater and hear the keys tap. I’m thinking of taking a stare-at-the-screen nap. When out in the server room there arose such a clatter. I jumped from my desk to see what was the matter. Away to the NOC I flew like a flash. Tore open the door and stroked my mustache.

The LEDs in the rack set the room all aglow, and gave a luster of blue and red, don’t you know? When what to my bloodshot eyes should appear, but a shower of sparks shooting out from the rear.

Out popped a fat man who ran out plenty quick, I knew in a moment it must be saint Nick. You bastard! I cried, and I called for my crew. Now, Mike!, Now Matt!, Now Jeremy!, Matt and Chicago Matt too! To the top of the building! Climb up that side wall! And get that fat jerk before his deer dash away all!

When we got to the top we had missed him by seconds, and the feeling in my gut was very much unpleasant. Back down in the server room the hard drives had been pulled, the tapes were all strewn out, and it was starting to flood.

That bastard had blown some metal filings in the air intake, then the power went out and it was all I could take. I went down to the bar and ordered a brew, then ordered 10 whiskeys and downed them all too.

When I came back to the office and threw up on the floor. The boss handed my my last paycheck and kicked me out the door. I got a new offer the very next day, making much more than double the pay. Thank you Santa, for this Christmas-time gift. Next year I’m definitely leaving you a tip.