Firstly, please allow me to make a few assumptions:

1. We want to run the backup every two hours and we want to keep 400 copies (around 17 days of data)
2. We want to backup some files on the local system….not the whole file system
3. We want to take snapshots of a database on the local system periodically
4. We will hold the backups on a remote server
5. This is a backup system of last resort, if we ever need to recover data from this system it will take manual intervention.

This system is run in two steps:

On the local server:

On the local server we want to dump the local mysql database to disk periodically, perhaps every two hours. To do this we need to follow these steps:

1. create a directory where we will be placing the database dumps
   /root/database_dumps
2. create a directory where we will place our scrips
   /root/scripts
3. save this file into the scripts directory
#!/bin/sh
# Script name : auto_mysql_dump.sh
# Backup the dbname database
date=`date +%F-%k.%M.%S`
DBS=`mysql -Bse ‘show databases’`
for db in $DBS
do
mysqldump –databases $db > /root/database_dumps/$db-$date.sql
done
cd /root/database_dumps
tar -czf mysqldump-$date.tar.gz *.sql
rm -fr *.sql
# delete all database entries that are more than 2 day old
find /root/database_dumps/ -amin +300 -exec /bin/rm -f {} \;

#!/bin/sh
# Script name : mysql_dump.sh
# Backup the dbname database
# Script created by Ross Cooney from Rozmic (www.emailcloud.com)

USER='your-mysql-username'
PASS='your-mysql-password' 

date=`date +%F-%H.%M.%S`
DBS=`mysql -u $USER -p$PASS -Bse 'show databases'`

for db in $DBS
do
       mysqldump -u $USER -p$PASS --databases $db > /root/database_dumps/$db-$date.sql
done

cd /root/database_dumps
tar -czf mysqldump-$date.tar.gz *.sql
rm -fr *.sql

# delete all database entries that are more than 3 hours old
find /root/database_dumps/ -amin +180 -exec /bin/rm -f {} \;

4. setup the following cron job:
   0 1,3,5,7,9,11,13,15,17,19,21,23 * * * /root/scripts/mysql_dump.sh > /dev/null

These four steps will setup a process that dumps the mysql database to disk every two hours and then hold a copy of that database dump for three hours.

On the remote server:

We want to run a script every two hours to copy the changes of any files on the disk. To do this you need to follow these steps:

1. create a directory where we will be placing the backup files
   /root/backup
2. Create a directory that we will use for this particular block of data, we use the IP address of the server but you can use anything that you want.
   /root/backup/10.0.0.6
3. create a directory where we will place our scrips
   /root/scripts
4. Ensure that the SSH key for the remote server is listed in the local server’s ‘authorized_keys’ file so that ssh connections can be made without needing a password.
5. Save this file to your scripts directory:

#!/bin/bash
# this script is called: 10.0.0.6.backup.sh
# Script created by Ross Cooney from Rozmic (www.emailcloud.com)

RSYNC_ARGS="-az --progress --delete --delete-excluded --delete-after"
BACKUP_DIRS="/var/www/html /etc /root/scripts /root/database_dumps"
SERVER="10.0.0.6"
DIR="/root/backup/10.0.0.6"

 rm -rf $DIR/daily.400

 for i in `seq 400 -1  1`; do
        let NEXT=$i+1
        mv $DIR/daily.$i $DIR/daily.$NEXT > /dev/null 2>&1
done

cp -al $DIR/daily.0 $DIR/daily.1

for RDIR in $BACKUP_DIRS; do
   rsync $RSYNC_ARGS ${SERVER}:$RDIR $DIR/$HOST/daily.0/
done

6. Setup the following cron job:
   0 0,2,4,6,8,10,12,14,16,18,20,22 * * * /root/scripts/backup.10.0.0.6.sh > /dev/null

These scripts work for us, I hope that they are of value to you, please dont take my word for their effectiveness, you should test them fully before you use them. I used the following resources when I was writing these scripts:

• http://www.mikerubel.org/computers/rsync_snapshots/
• http://bash.cyberciti.biz/backup/backup-mysql-database-server-2/

Update (2nd October 2009):
Thanks to Oli Wood for pointing out the typo in the above script, this has now been resolved

The speakers this month will be:

• Jonathan Noble — Microsoft Windows PowerShell, an introduction
  Microsoft released Windows PowerShell in 2006 to finally give Windows a decent shell and scripting language for automation and PowerShell V2 will come installed and on by default in the forthcoming releases of Windows 7 and Server 2008 R2. Jonathan will give an overview of PowerShell and explain why it’s not just of interest to Windows sys admins.
• Jonathan Bradford: Online, Offline?
  A new generation of applications are emerging which will bring “online” applications to “offline” situations.  What might they look like? And how will they be built?
• Mike Parker – Rapid web app development using Drupal as the Framework.
  Mike will give an insight into how quickly you can build an app and how you may go about doing it.

The event is in the Beehive at Newcastle University and starts at 18:00.

I think ‘cost’ for a student is measured in £££, which is important, but, as a business owner and a developer I think that time often is more important than £££. 

Open source software is a low cost alternative to proprietary software.  For example, the Linux operating system is a low cost alternative to the Microsoft Server 2003 operating system. This popularity is very evident when you see that today the largest market share for web servers is held by the open source Apache system (51% market share). Many think that the key advantage of open source software is its low cost of ownership. I believe that this is not the case. 

As a business owner I see other advantages, namely: 

• Lower total cost of ownership 
• Reduced dependence on software vendors 
• Easier to customize 
• Higher level of security 

By far, the most important thing that open source software gives me is independence. I really value my vendor independence. This includes independence from the vendor for maintenance, support and the 
necessity accept version upgrades that I do not want or need. 

For example, when Microsoft releases a new operating system it usually phases out support for older versions. This would force me to upgrade the OS on our network, even if we felt that we didn’t need the update. In contrast, there is no forced upgrade cycle with open source. Older versions of open source products continue to be supported through the open source community and third party support providers as long as there is demand in the marketplace for such support. 

I am not a Microsoft basher, we do use Microsoft products on our network. We use WinXP on all internal desktops with Office and Micorsoft Exchange on our internal network. But it would be very unlikely that I would use it on may of our scan network.