Ruby if – else:
http://weblog.jamisbuck.org/2007/1/5/brevity-vs-clarity

Capistrano custom maintenance page:
http://clarkware.com/cgi/blosxom

Rails Rubyisims:
http://errtheblog.com/post/44

1. DO NOT have ruby logger rotate itself. In a clustered environment all instances will try and rotate the log at once. One will win and stay alive all others will crap out. Have an external program manage the logs for you.
2. If using capistrano change the svn setting from checkout to export:
   set :checkout, “export”
   This is for security reasons because there is extra details in the .svn metadata in a checkout opposed to an export.
3. Set the ActiveRecord database verification timeout:ActiveRecord::Base.verification_timeout = 14400Place this in your environment.rb right after the end of Rails::Initializer.run. This should be seet lower then interactive_timeout for your database. In a mysql prompt you can type
   SHOW VARIABLES;
   to find the value. In mine it was set to 28800.
4. Create a maintenance page. This is easily managed thru capistrano and in my situation allowed me to direct all traffic to the maintenance page while I left the server running and debugged it locally. It will come in handy.
5. Use the Exception Notifier plugin and clean up the 404 and 500 html files in the public dir. The plugin uses them and sends you the error in an email.
6. CAPTURE DATA! I am going back and modifying several tables to capture more data. Rails convenient created_on and other helpers are excellent, use them. You can never collect too much data, we are in the information age.

SYMPTOM:
Late in the evening all instances of mongrel but one would return a Status: 500 Internal Server Error. One instance would still work fine, strange enough it was always the second instance started that would still be alive. If you ran one instance it was fine. The error would present itself only in a mongrel clustered environment.

Thanks to Eden on the mongrel mailing list, the solution was found:
http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/218107

SOLUTION:
DO NOT allow the ruby logger to rotate itself in a clustered environment. Have some other external program manage the logs for you.

Create a file .my.cnf in your home directory. Put the following entry into it:

[mysql]
user=user
password=pass

Then create a file in /usr/local/bin called db or whatever, and put this in it:

#!/bin/bash
mysql db_name

Now from anywhere in the terminal you can type . db and automatically have a mysql prompt. Lovely!

We had some issues with email coming from the MyFreeCopyright.com server being flagged as spam from Yahoo, Hotmail, and AOL. Give thanks that I have a super linux guru on the team. A couple tweaks and everything is all good. If you need this or another trick done to fix your companies emails from being flagged as spam, let me know.

After looking into the apache error log I saw this:

proxy: error reading status line from remote server ...

After doing some googleing I found this article.

It states that the following lines need to be placed in your apache conf to fix a current bug in mod_proxy:

#Fix for Apache bug 39499
SetEnv force-proxy-request-1.0 1
SetEnv proxy-nokeepalive 1

The change has been submited to my conf, I will post the results.

The line looks like this:

RAILS_DEFAULT_LOGGER = Logger.new("#{RAILS_ROOT}/log/#{RAILS_ENV}.log", "daily")

You can find other parameters to pass to it by looking at the ruby Logger API.

I must now go and consume this bubbly.