What does your backup system do?

Let's assume everyone has a backup system that runs automatically and does everything right every time. Are you with me so far? This little assumptions covers a lot, so it's ok to spend some time thinking about it. If you have a system like this it actually means

• that you know where are of your important data is stored,
• that the location(s) your important data is stored
  • never changes or
  • changes and your backup system is informed about it,
• that you get a daily report (you want to know if it's working, right?),
• that your storage space never fails and has enough room to accommodate growth.

If you are unsure whether your backup system has all of these properties, you better check. In case your backup system doesn't run unattended and with defined periods you might not have a backup system at all. Designing backups and regularly inspecting existing backup infrastructure is well spent time. Detecting that the backup system failed long ago (but no one noticed) is something that should never happen. Should.

There are plenty of strategies to explore when creating an automated backup system. The most important part is a reliable transfer of the data and proper storage. Especially a backup server must have redundant disks. You don't need to run a high-availability cluster, but your data must be stored safely. Recreating a full backup if your live systems cover hundreds of gigabytes or terabytes is not an option. Above all select tools you know how to operate. If you can't cope with the CLI or GUI, then you are using the wrong tools.

Recovery is even more important than backups. If you have backups, then do a recovery. Do it every week, month, quarter or year, but do it. Recover a full workstation or server on spare hardware (or clouds) just to see if it works. Do it now!

• Amanda Open Source Backup