Backup catalog

A backup catalog is the database that backup software keeps about every restore point: what it contains, when it ran, where the data is stored, which earlier backups it depends on and when it expires. Every restore starts with a catalog lookup, so intact backups are slow or impossible to use once the catalog is gone.

What does a backup catalog record?

Each restore point gets one catalog record, which typically holds these fields:

  • Contents: the files, databases or virtual machine disks captured, often down to individual file paths.
  • Time: when the job ran, which defines the moment a restore returns to.
  • Location: where the data sits on the backup target, down to the specific files or objects.
  • Dependencies: the chain of earlier backups needed to rebuild it, such as the full behind an incremental or differential.
  • Expiry: the date, set by the retention policy, after which the restore point can be removed.
  • Encryption details: whether the data is encrypted and which key or password opens it.

A request for last Tuesday's finance share is resolved by the catalog into one restore point and a list of objects to read. Only then does the software touch storage.

How does the catalog control deletion?

The catalog is also the authority for deleting. When retention passes, the backup software checks the catalog to find which restore points have expired, and the dependency field keeps a full backup in place until every incremental or differential built on it has expired too. A catalog that has been corrupted or edited can therefore mark live restore points as expired, and the software then deletes them itself.

What happens to the backups if the catalog is lost?

The data on the target survives, but it can no longer be found by name. Most products can rebuild a catalog by scanning the stored backup files and reading the metadata written alongside them. On a large repository that rescan can take hours or days, and no restore is possible until it finishes. If the backups were protected by backup encryption with a password held only in the lost configuration, the rescan has nothing to open them with: the files are intact but unreadable.

How is a catalog kept recoverable?

Catalog protection rests on three measures. The backup server's configuration and catalog are backed up on their own schedule to a location separate from the repository they describe. Encryption passwords are stored away from the backup server, in a password vault or sealed record that stays reachable when the server is gone. Catalog imports are rehearsed on a clean server, which also measures how long a rebuild takes at current repository size, a figure that sits in the recovery plan next to restore throughput.

Why is the catalog an attack target?

The catalog usually sits on the backup server, so an attack on that server often takes out the catalog along with the path to the data. An attacker who cannot delete locked backups can still make them hard to use by damaging the index, the encryption material or the retention settings. Protecting the catalog is part of protecting against backup tampering.

Frequently asked questions

Is the backup catalog the same as the backup repository?

No. The repository holds the backup data. The catalog is the index that tells the software which data belongs to which restore point.

Can backups be restored without the original catalog?

Usually, by importing or rescanning the stored backup files into a new catalog. It takes longer than a normal restore and needs any encryption passwords.

How big is a catalog?

It depends on the number of files and restore points, and file-level catalogs for large file servers can be significant. Size matters less than where it is stored and whether a second copy exists.