What is the 3-2-1 backup rule?
The 3-2-1 backup rule is a simple plan for not losing data: keep three copies of anything important, store them on two different types of storage, and keep one copy off-site. The three copies are the live data plus two backups, so one failure, mistake or disaster never takes out every copy at once.
Think of your house keys: you carry one set, a second set hangs in a drawer at home, and a neighbour across town holds a third. Losing your bag costs one set. A burglar who takes the bag and the drawer set still leaves you the neighbour’s keys.
What does each number mean?
3 copies
The original counts as one, so you need two backups on top of it. Two backups means a failed backup job or a corrupt file still leaves you a good copy to restore from.
2 types of storage
Copies on the same kind of storage share the same weak spots. A batch of faulty disks, a file system bug or one storage platform’s outage can hit both. Mixing types, such as local disks and object storage, or disk and tape, keeps one fault class from reaching every copy.
1 copy off-site
One copy lives somewhere a fire, flood, power loss or theft at the main site cannot reach. For cloud workloads, off-site means a different provider or at least a different zone and account from the live data.
What does not count as a copy?
Redundancy inside one system protects against hardware failure, not against the mistakes that backups exist for. If someone deletes a bucket or ransomware encrypts a volume, every replica changes at the same moment.
- RAID, replication and erasure coding inside one storage system are a single copy.
- A snapshot on the same volume or cluster is a fast undo, but it shares the system’s fate.
- A sync job that mirrors deletes to the second location copies the damage too.
- A backup that has never been restored is a guess, not a copy you can rely on.
What is 3-2-1-1-0?
3-2-1-1-0 is a common extension of the rule. The extra 1 asks for one copy that is immutable or offline, so an attacker with admin access cannot change or delete it. The 0 means zero errors when you test a restore.
Ransomware often goes after backups first, because a working backup makes the ransom pointless. An immutable copy closes that gap. Regular restore tests prove the copies are complete, and they tell you how long a real recovery takes, which feeds your RPO and RTO.
How enum handles the 3-2-1 backup rule
enum Object Storage can hold the off-site or second-media copy for backup tools that speak the S3 API. Versioning recovers overwritten or deleted objects, Object Lock keeps backups immutable for the retention you set, and lifecycle rules expire old versions. Storage runs in one zone without bucket replication, so keep a further copy elsewhere for anything you cannot afford to lose.
FAQ
Does the live data count as one of the three copies? Yes. In the original rule the three copies are the primary data plus two backups. Some teams count three backups on top of the original, which gives more margin but costs more storage.
Is a cloud bucket enough to meet the off-site part? It can be, as long as the bucket sits outside the failure zone and admin accounts of the live data. A bucket in the same account and zone as production shares too much with it.
How often should I test a restore? Test on a fixed schedule and after every change to the backup setup. Restore both single files and full systems, and roll back several days, so you know older copies work and how long recovery takes.