You have a backup. But can you restore it?

The most expensive lesson learned in a disaster is that the backup has been empty or corrupt for months. When backup software says “successful”, it does not show that the files open, that the database is consistent or that the restore will fit into a working day.

The 3-2-1 rule still holds

• 3 copies: the original data and two backups.
• 2 different media: for example local storage and the cloud.
• 1 off-site copy: fire, theft and flood take the backup in the same building too.

Ransomware now looks for the backups first and encrypts them. So we add one more item to the rule: one of the copies must be immutable, and must not be deletable with an administrator account on the network.

Set two numbers: RTO and RPO

RTO is the number of hours within which a system must be back up; RPO is how many hours of data loss you can tolerate. Four hours may be acceptable for the accounting server, while two days may be enough for the archive. The backup architecture is built around these two numbers; without them, the investment is either too little or too much.

Without a restore test you cannot be sure

For the restore test, once a quarter we restore a randomly chosen server and a few files from backup, measure the time and report the result. That report is also one of the pieces of evidence most often requested in audits and bank questionnaires. See our Backup and Disaster Recovery page for details.