I recently got a new VPS from vpslink.com. Their service is great because I can choose from a list of operating systems. I, of course, chose “Hardy Heron”, Ubuntu 8.04. This service is totally unmanaged other than DNS and billing so I have to manage everything myself, including backup.

I haven’t found anything out there that I’ve fallen in love with as far as backup is concerned, so I started with a shell script that covered the basics. I needed a rotating incremental backup script with full backups performed once a week. I wanted all the backups to live in a top level path like /backup and I wanted the whole backup process to be owned and operated by a backup user. My distribution came with a backup user, it just needed to be configured.

So I created a home directory for the user at /home/backup and set its ownership and also created a /backup folder and set its ownership.

sudo mkdir /home/backup
sudo chown backup:backup /home/backup
sudo mkdir /backup
sudo chown backup:backup /backup

The backup user will be using tar to backup my system, including files that even the backup user shouldn’t be able to directly modify. So I modified my sudoers file to allow the backup user to use tar without providing a password (so it can be done automatically).

Continue reading »

Backup seems like a very simple concept. The idea is to make an extra copy of important data so in the event of a failure or disaster, that data can be promptly recovered. But anyone who’s been through the data recovery process knows that its just not that simple.

There are a staggaring array of different choices to make when designing a backup solution. The decisions you make will be determined by the type of data and amount you need to back up, your reliability and downtime tolerances, and your budget. A good backup solution can be as simple as a USB thumb drive to hold backup accounting records or as complex as the most advanced disk to disk to disk solutions for backing up legacy database systems.

Continue reading »