What is Veeam immutable backup?
A Veeam immutable backup is a restore point that the storage underneath Veeam refuses to change or delete until a set date. Veeam asks for the lock, the storage enforces it, and the Veeam server cannot override it, which moves deletion power from the console to the storage.
Who asks for the lock and who enforces it?
Veeam sets a retention date on each restore point as it writes it. The repository enforces that date. On object storage the mechanism is S3 Object Lock. On a Linux hardened repository it is a file attribute that blocks changes. After the lock is set, a delete from the Veeam console, even from an administrator account, is refused by the storage. The security benefit comes from the split: a stolen Veeam login can still reach the console but no longer controls deletion.
Object storage supports two lock modes. In governance mode, a user with a special permission can remove the lock early. In compliance mode, no one can until the date passes. A lock that a privileged login on the storage can lift is protection against mistakes more than against attackers, so compliance mode is the setting that matches an immutable backup threat model.
How long does the lock have to last?
The lock has to outlast two things. The first is retention, since a restore point that unlocks before it is no longer needed can be deleted early. The second is dwell time, the period an attacker spends inside before acting. Where intruders sit in a network for weeks, restore points locked for days may all have been created after the intrusion began, and every recent copy could contain the attacker's tools or already-damaged data. Setting the period means asking how long it takes to notice an intrusion, find an older clean point and restore it. The stages are described under the ransomware kill chain.
Longer locks cost capacity, because data cannot be freed while it is locked. The balance is between recovery needs and what the repository can hold.
Why do locks outlast the stated retention?
Veeam stores data as blocks that several restore points can share. A block written for today's point may also be needed by a point from last week, so protecting the old point means protecting the shared blocks until the newest point that uses them expires. Veeam adds a block generation period to the immutability date on object storage for this reason, a number of days that varies by storage type. The result is that the locked footprint is larger than retention alone suggests, which matters when capacity is sized. The storage model is explained under Veeam backup to object storage.
What does the lock not stop?
- Stopped jobs. A hijacked console cannot delete locked restore points, but it can disable jobs or change schedules, so that no new points are made. Alerts on job failure and missing restore points are the check on that.
- Reading. Locked files can be read and copied, so immutability does not prevent data exfiltration.
- Bad data. A locked copy preserves whatever was backed up. If files were already damaged when the job ran, the locked point holds the damage.
- The lock period itself. A period shortened in the console affects only new restore points, so older locks hold, but each new point will be locked for less time.
Frequently asked questions
Is a Veeam immutable backup the same as an offline backup?
No. It stays online and reachable, but cannot be altered or deleted for the lock period. An offline backup is disconnected from the network.
Can an administrator delete an immutable restore point?
Not through Veeam. Whether anyone can remove it depends on the storage mode: in compliance mode no one can before the date, while in governance mode a privileged storage user can.
Does immutability apply to the configuration backup?
Immutability applies to restore points in repositories that support it. The configuration backup is a separate item, and its protection depends on where it is stored.
