Monit: http://www.tildeslash.com/monit/

Munin: http://munin.projects.linpro.no/

GOD: http://god.rubyforge.org

Of the 3 above solutions, i have found GOD to be the easiest to set up. I will try and outline the steps i took to get things working. I will also show you how i integrate this with Google Apps Mail for email notifications. The details of how to manipulate the GOD configuration are explained in depth on the GOD website, so i will not discuss that here.

My production server is running the following:

• Rails 2.0.2
• Ubuntu Gutsy
• nginx 0.5.26
• mongrel 1.1.4
• mongrel_cluster 1.0.5
• god 0.7.8
• tlsmail 0.0.1

Step 1:

Install the god gem:

sudo gem install god

Step 1a (OPTIONAL):

Install the tlsmail gem if you want to use GMail:

sudo gem install tlsmail

You can read more about using GMail in your rails app in my previous blog post:

http://therailscoder.typepad.com/the_rails_coder_ruby_on_r/2008/04/using-gmail-smt.html

Step 2:

Find and edit the init.d script.

cd /etc/init.d
ls -l
sudo nano /etc/init.d./god

It should look like this:

http://pastie.org/250816

Run these commands to make the god script executable and allow it to start on a reboot of the server.

chmod +x /etc/init.d/god
/usr/sbin/update-rc.d god defaults

Step 3:

Create your god config file.

cd /etc
sudo nano god.conf

This is basically a configuration taken from the god website, but i had to make some changes to the location of the pid files, make sure you get this right – otherwise things will not work the way they should (you can also exclude the code from line 60 onwards if you are not using email notifications with GMail):

http://pastie.org/250821

Step 4:

Start god:

sudo /etc/init.d/god start

You should see something like this:

Sending 'load' command

The following tasks were affected:

mydomain-mongrel-8000

mydomain-mongrel-8001

mydomain-mongrel-8002

Step 5:

Check whether your mongrels are being monitored:

sudo god status

You should see something like this:

mydomain-mongrel-8000: up

mydomain-mongrel-8001: up

mydomain-mongrel-8002: up

If god is not monitoring your mongrels, you will see this:

mydomain-mongrel-8000: unmonitored

mydomain-mongrel-8001: unmonitored

mydomain-mongrel-8002: unmonitored

Step 6:

Now, to really see god in action, i look at the log files:

sudo god log mydomain-mongrel-8000

sudo god log mydomain-mongrel-8001

sudo god log mydomain-mongrel-8002

You should start seeing messages like this, which means it is working:

INFO: mydomain-mongrel-8000 [ok] process is running (ProcessRunning)
INFO: mydomain-mongrel-8000 [ok] memory within bounds [59456kb, 59456kb, 59456kb, 59456kb, 59456kb] (MemoryUsage)
INFO: mydomain-mongrel-8000 [ok] cpu within bounds [0.00891569765038892%, 0.00891040849797263%, 0.00890512561731393%, 0.00889984899726403%, 0.00889457862670062%] (CpuUsage)

Step 7:

To stop god:

sudo /etc/init.d/god stop

Step 8:

Lastly, I recommend changing the memory or cpu useage to a lower level and then watching the logs to see that god is picking up on this and restarting the mongrels, as well as sending emails!

Conclusion:

That's it, check out the GOD website and Google Group for more info!

mysql -u root -p

Other commands;

show databases;
use enter_database_name_here;
show tables;
describe enter_table_name_here;

Show data from a particular column in a table;

SELECT enter_column_name_here FROM enter_table_name_here;

Update the data in a particular column in a table;

UPDATE enter_table_name_here SET enter_column_name_here = ‘enter_new_value_here’ WHERE id =’enter_id_here’;

My method now is to design first for Firefox and Safari. If things look good on these browsers, they are likely to be good on Opera as well. Then, i make tweaks to fit IE7, i’ve not yet started developing for IE8 as it is still in beta. IE7 has its quirks, but it is not too difficult to get things looking similar or exactly the same as it would on Firefox and Safari. IE6 is another story all together. It is painful trying to get CSS to look the way you want on this browser. IE6 has a number of bugs and ‘bug fixes’ or hacks that are well documented on the web.

But having said that, if you want to adjust your website design to fit Internet Explorer whilst on a Mac, without having to move over to a Windows PC – there are a few options that could make things easier.

Option 1:

Use Parallels.

http://www.parallels.com

Parallels basically lets you install a version of Windows on your Mac. You can then install and use any Windows software on your Mac. Windows XP is probably a better bet than Windows Vista. The reason being that there are some known issues getting IE6 to work on Vista. If you install XP, you will get IE6 as standard, but you can upgrade to IE7. You can then install the Multiple_IE software which will give you IE versions 3 to 6:

http://tredosoft.com/Multiple_IE

You could probably also use Firefox and get a multiple IE plugin (these plugins need Windows, and won’t work on Mac Firefox):

https://addons.mozilla.org/en-US/firefox/addon/1419

Parallels costs about US$79.99.

Option 2:

UseVMware Fusion.

http://www.vmware.com/products/fusion/

This lets you run a ‘Virtual PC’ on your Mac, i don’t know much about this, but cost is also about US$79.99

Option 3:

Use ie4osx.

http://www.kronenberg.org/ies4osx/

This lets you use IE5 to 7. It does not require Parallels or VMware Fusion. It is FREE. And it is dead simple to install. Unforunately it is supposed to be a bit buggy with IE7, but hopefully that will change soon. This is the option i chose as IE6 is what i need to test the most, and is the hardest to code right.

Hope this helps someone.

Anyway, add the “show_simple_captcha” code to your new user signup form and include the parameter bject => ‘user’. Then add the following to the user model – models/user.rb:

apply_simple_captcha :message => “The text does not match the image”, :add_to_base => true
attr_accessible :login, :email, :password, :password_confirmation, :identity_url, :profile, :captcha, :captcha_key

And finally change the user_controller.rb create action:

def create
cookies.delete :auth_token
@user = User.new(params[:user])
if @user.save_with_captcha
#Uncomment to have the user logged in after creating an account – Not Recommended
#self.current_user = @user
flash[:notice] = “Thanks for signing up!”
redirect_to login_path
else
flash[:error] = “There was a problem creating your account.”
render :action => ‘new’
end
end

That’s it.

http://blog.pomozov.info/posts/how-to-send-actionmailer-mails-to-gmailcom.html

Stephen Chu talked about how to use GMail’s smtp server back in 2006:

http://www.stephenchu.com/2006/06/how-to-use-gmail-smtp-server-to-send.html

More recently, Preston Lee talks about the Anatol Pomozov approach here:

http://www.prestonlee.com/archives/63

And Daniel Fischer explains how to use the action_mailer_tls plugin – this method is also mentioned in the new Advanced Rails Recipes book:

http://www.danielfischer.com/2008/01/09/how-to-use-gmail-as-your-mail-server-for-rails/

However, there is another far easier solution that was posted originally to my knowledge by Tim Riendeau:

http://www.wanlord.com/articles/2007/11/29/sending-email-using-actionmailer-and-gmail

I will repeat the steps as they worked for me here…

1. Install the tlsmail gem

sudo gem install tlsmail

2. Add the following to the bottom of your environment.rb

require ‘tlsmail’
Net::SMTP.enable_tls(OpenSSL::SSL::VERIFY_NONE)

ActionMailer::Base.delivery_method = :smtp
ActionMailer::Base.smtp_settings = {
:address => ‘smtp.gmail.com’,
:tls => ‘true’,
:port => 587,
:domain => ‘mydomain.com’,
:authentication => :plain,
:user_name => ‘info@mydomain.com’,
:password => ‘mypassword’
}

3. To enable errors to show up if the mail does not get delivered, modify your development.rb and production.rb as shown below – it’s ok to make this true in devlopment, but i think most people say to set this to false in production:

development.rb
config.action_mailer.raise_delivery_errors = true

production.rb
config.action_mailer.raise_delivery_errors = false

And that’s it!

NOTE: “Timeout::Error (execution expired)”

I got this message quite a lot when i tried to send emails in development, and no one seems to know why this is the case. But when i switched to production things worked like a breeze. I am developing on a MacBook Pro with Leopard, if you have any idea why this happens please let me know.

NOTE: Internet Service Provider blocking ports

Another point to note is that some ISP’s block outgoing smtp mail on certain ports, eg port 25. Check with your ISP if they do any port filtering.

Now there are a lot of options available to achieve all this, this is the way that i found easy and that worked for me. To create the automatic mysql database dumps, there is the automysqlbackup script available from SourceForge:

http://sourceforge.net/projects/automysqlbackup/

I downloaded this onto my Macbook Pro and used ‘scp’ to copy back to my server. I suppose you could use ‘wget’ directly form the server too if you prefer. You can put the file anywhere on the server you like, i just keep it in my home directory for now (/home/user/) – but i will move it later. First thing is to rename the file, at the time of writing version 2.5 was available:

sudo mv automysqlbackup.sh.2.5 automysqlbackup.sh

Now create a directory in which you want your backups to go to, and one where the restored backups from S3 will be downloaded to, eg /home/user/mysqlbackups/ and /home/user/mysqlrestores/:

mkdir mysqlbackups
mkdir mysqlrestores

The automysqlbackup script is of course just one file, and you can open and view it in a text editor. In fact i suggest that you do that next as it has pretty good instructions as to what it does and what parameters you need to add. Move to the directory in which you have the script and open it up, i use the nano editor:

sudo nano automysqlbackup.sh

Basically you need to add the names of your databases, mysql user, and mysql password info. The script will do daily, weekly, and monthly backups of your database into the specified backup folder of your choice – you can even have emails sent to you! Here is an excerpt:

# Username to access the MySQL server e.g. dbuser
USERNAME=root

# Backup directory location e.g /backups
BACKUPDIR=”/home/user/mysqlbackups”

Edit the files with your own settings, then save and exit!

Next edit the permissions on the script to give read, write, and execute permissions to the user, and not to to anyone else.

sudo chmod 700 automysqbackup.sh

You can test the script out manually as follows:

./automysqlbackup.sh

Move to your backups directory and you should see 3 sub-directories – daily, weekly, and monthly. In the daily directory you will find a gzipped sql file of you database backup! I will talk about how to actually automate this process a little later on…

Next we need to get the mysql backup’s over to Amazon S3. If you haven’t got an S3 account, sign up and get your Access Key ID and Secret Access Key ready:

http://www.amazon.com/s3

Next we need to setup s3sync:

http://s3sync.net

I want to credit John Eberly’s blog post from way back in 2006 for helping me set up s3sync. I have cheekily repeated most of his instructions in some of the steps below…

http://blog.eberly.org/2006/10/09/how-automate-your-backup-to-amazon-s3-using-s3sync/

s3sync requires ruby 1.8.4 or greater as well as libopenssl-ruby. I already had these setup on my server so i did not have to install them, if you don’t have these enter the command below to get them:

sudo aptitude install ruby libopenssl-ruby

In your /home/user/ directory download and install s3sync as follws:

wget http://s3.amazonaws.com/ServEdge_pub/s3sync/s3sync.tar.gz
tar xvzf s3sync.tar.gz
rm s3sync.tar.gz

You then need to create a directory for ssl certificates as well as download them. Move to the newly created s3sync folder and follow the commands below:

cd s3sync
mkdir certs
cd certs
wget http://mirbsd.mirsolutions.de/cvs.cgi/~checkout~/src/etc/ssl.certs.shar

Run this command form the certs directory:

sh ssl.certs.shar

Now go back to the main s3sync directory. You can create an S3 ‘bucket’ to store your database backup using s3sync eg:

ruby s3cmd.rb createbucket mysql.mydomain.com

Or use an GUI interface such as S3 Browser, Cockpit, or Jungle Disk:

http://people.no-distance.net/ol/software/s3/
http://jets3t.s3.amazonaws.com/applications/cockpit.html
http://www.jungledisk.com/

Next create your upload and download scripts, and edit them as follows with your own parameters. Note the ‘database’ is just a name i gave as a sub-directory or folder within the bucket – you can call it what you want:

sudo nano upload.sh

#!/bin/bash
# script to upload local directory upto s3
cd /home/user/s3sync/
export AWS_ACCESS_KEY_ID=yourS3accesskey
export AWS_SECRET_ACCESS_KEY=yourS3secretkey
export SSL_CERT_DIR=/home/user/s3sync/certs
ruby s3sync.rb -r –ssl –delete /home/user/mysqlbackups/ mysql.mydomain.com:/database
# copy and modify line above for each additional folder to be synced

sudo nano download.sh

#!/bin/bash
# script to download local directory upto s3
cd /home/user/s3sync/
export AWS_ACCESS_KEY_ID=yourS3accesskey
export AWS_SECRET_ACCESS_KEY=yourS3secretkey
export SSL_CERT_DIR=/home/user/s3sync/certs
ruby s3sync.rb -r –ssl –delete mysql.mydomain.com:/database/ /home/user/mysqlrestores
# copy and modify line above for each additional folder to be synced

Change the permissions on the files, and check the ownership – i had to change this to the current user (me):

sudo chmod 700 upload.sh
sudo chmod 700 download.sh

sudo chown user upload.sh
sudo chown user download.sh

Set up some S3 configuration:

mkdir /etc/s3conf
cp s3config.yml.example /etc/s3conf/s3config.yml

Edit s3config.yml with your S3 details and ssl cert path:

sudo nano /etc/s3conf/s3config.yml

In your s3sync directory, to run the upload and download scripts manually enter:

./upload.sh
./download.sh

To check the upload use ruby scmd.rb as shown below, or use your preferred GUI software:

ruby s3cmd.rb list mysql.mydomain.com

To check the download just go to your mysqlrestores directory. I found that when i tried the download, i needed to also create a sub-directory called ‘database’ (/home/user/mysqlrestores/database/).

What we now have is the ability to create mysql database backup files, as well as being able to upload and download these backup files to and from S3. But everything is being done manually! I’ll finish up by showing how this can become an automated process…

To get the automysqlbackup script to function automatically there are a couple of options. On Ubuntu and most other Linux distributions you should have the following ‘cron’ directories:

/etc/cron.daily/
/etc/cron.weekly/
/etc/cron.monthly/

You can move the script to any one of these directories, and the script will run automatically every day, week, or month! On Ubuntu the .sh extension does not need to be there.

sudo mv automysqlbackup.sh /etc/cron.daily/automysqlbackup

But i prefer not to do this! This is because there are other default scripts or processes that run on a daily basis, and these will be run at the same time as your database backups. The default on my setup is 0625 am! I prefer to have my database dumps run at a different time, when there are no other scripts or processes running, and when my website is likely to have less traffic. The same goes for uploads to S3. The easiest way to do this is to create your own ‘crontab’, here you can specify what time to run your scripts.

Before i do this, i like to move the automyslbackup script to the s3sync directory, where it’s friends ‘upload’ and ‘download’ live !

sudo mv automysqlbackup.sh /home/user/s3sync/automysqlbackup.sh

Then from any directory open your own ‘crontab’:

crontab -e

Then paste in this code:

# m h dom mon dow command
0 1 * * * /home/user/s3sync/automysqlbackup.sh
30 1 * * * /home/user/s3sync/upload.sh

Save and exit. This will basically set things up so that the automysqlbackup script will run every day at 0100 am, and so that the s3sync uploads the backups to S3 at 0130 am every day! You might need to do a Google search on crontab to find out how to customize this to suit your own needs.

Done!

Step 1.

On my Ubuntu Gutsy server, i create a mysqlbackups directory and then do a mysqldump into this directory:

mysqldump -u root -p databasename > databasebackup.sql

Step 2.

On my Macbook Pro, i create a mysqlbackups directory and the copy the databasebackup.sql file from the server to my local machine:

scp user@11.222.333.44:/home/user/mysqlbackups/databasebackup.sql /Users/user/mysqlbackups

Step 3.

Upload the databasebackup.sql file to an Amazon S3 bucket using S3 Browser (see my previous post for more info on this).

AND/OR

Backup to CD/DVD/USB/External Hard Disk.

This is obviously a manual process that could become quite tedious. If anyone has any suggestions as to how to backup MySQL databases to Amazon S3 that is more automated, please let me know how you would do this!

So how do we implement it?

First create an Amazon Web Services account, and sign up to use S3:

https://aws.amazon.com

Note that you will have to give some credit card details as the service is not free, but it is damn cheap!!

You will be given an access key id and a secret access key. These are both very long strings.

Next step is to create a bucket. A bucket is like a folder that will hold all your files. The easiest way to do this is get an S3 GUI, just like CocoaMSQL(Mac)/HeidiSQL(Win) for MySQL. There are a number of options to consider, eg:

S3 Browser:

http://people.no-distance.net/ol/software/s3/

Cockpit:

http://jets3t.s3.amazonaws.com/applications/cockpit.html

JungleDisk

http://www.jungledisk.com/

As i am using a Macbook Pro, i opted to use the S3 Browser.

With the S3 Browser, you just need to create a New Connection by entering you access key id and secret access key. There will be a similar option for other S3 GUI’s. Once you’ve done this you will be connected to your S3 storage and can simply add and delete buckets as you wish! I think it is a good idea to have a separate bucket for the development, test, and production versions of your app (just like you do with databases). With the S3 Browser you can also view the contents of your buckets and delete files from the buckets as well.

Next step is to install the S3 gem, eg:

sudo gem install aws-s3

Now get back to your app. I use attachment_fu for photos. So in my Photo model, i have the following:

has_attachment :content_type => :image,
:storage => :s3,
:size => 0.megabyte..3.megabytes,
:resize_to => ’450×450>’,
:thumbnails => { :thumb => ’150×150>’ }

validates_as_attachment

Go to your Rails app config directory and you should see the amazon_s3.yml file.

Edit it as follows. Basically you need to enter your access key id and secret access key’s plus specify the bucket names:

development:
bucket_name: photos_myproject_development
access_key_id: xxxxxxx
secret_access_key: xxxxxxxxxxxx

production:
bucket_name: photos_myproject_production
access_key_id: xxxxxx
secret_access_key: xxxxxxxxxxxx

I am using Subversion for source control, but i do not check the amazon_s3.yml file into Subversion, in fact i treat it just like database.yml (there are different ways of dealing with this, i won’t be going into this right now).

That’s it. Pretty easy. Upload your files and then use your S3 browser to see that they are indeed on S3!

http://www.google.com/a/