RAID Is Not a Backup: What a Two-Drive Mirror Protects Against, What It Doesn't Protect Against at All and Why Server and NAS Owners Lose Data While Sure That "Everything Is Duplicated"

Offices, studios and advanced home users often have a server or network storage device with several drives combined into a RAID array. The most popular option is a two-drive mirror, where every file is written to both drives at once. That leads to a belief: "the data is duplicated, so we don't need a backup". It's one of the most dangerous misconceptions in data management. RAID really does protect against one specific disaster — a drive failing. But it doesn't protect against most other causes of data loss at all, and in some cases even spreads them faster.

What RAID is and why it's needed

• RAID 1 (mirror) — two drives hold identical data. If one dies, work continues on the other.
• RAID 5 — three or more drives, with data spread out along with parity information. It survives one drive failing.
• RAID 6 — like RAID 5, but survives two drives failing at once.
• RAID 10 — a combination of mirroring and striping: fast and reliable, but needs at least four drives.

RAID's main job is keeping things running. When a drive fails, the server doesn't stop, staff keep working, and the faulty drive gets replaced. That's valuable, but it isn't a backup.

What RAID doesn't protect against

• Deleted files. An employee deletes a folder — it's instantly deleted on every drive in the array.
• Ransomware. The virus encrypts files, and the array dutifully writes the encrypted versions to every drive.
• Software and user errors. A corrupted 1C database, an overwritten file, a failed update — the damage is duplicated.
• Failure of the controller or the NAS itself. If the device managing the array dies, the drives may be unreadable without an identical controller.
• Power surges, fire, flooding, theft. All the array's drives sit in one box and suffer together.
• Several drives failing at once — more than the array can tolerate.

Insidious scenarios

A degraded array nobody noticed. One drive in a mirror failed six months ago. The server keeps working; the warning blinks on a light in the server room or goes to an inbox nobody reads. When the second drive dies, the data is lost completely.

Failure during rebuild. After a drive is replaced, the array starts rebuilding — reading all the data from the remaining drives for hours. That load often finishes off a second, already worn drive, especially if all the drives were bought at the same time from the same batch.

Wrong moves during a failure. The wrong drive replaced, the array re-initialised, a rebuild "in the wrong direction" — and the working data is overwritten with data from an old drive.

The right way

RAID and backups solve different problems and must work together. RAID makes sure one drive failing doesn't stop work. A backup makes sure that any other problem doesn't lead to permanent loss. The right setup for an office: a server or NAS with RAID, plus regular automatic backups to a separate device, plus a copy off-site or in the cloud. Backups should keep version history so you can go back to the state before an infection or mistake, and be protected from being altered by ransomware. Monitoring is also essential: drive health alerts must reach a responsible person who acts on them.

If the array has already failed

If the RAID is no longer recognised, has fallen apart, or the NAS reports errors on several drives, the key thing is not to try to "rebuild" the array, initialise it or swap drives at random. Each of those actions can irreversibly overwrite the array's metadata. Number the drives in their installed order, switch the device off and contact specialists. RAID recovery usually starts by imaging each drive and assembling the array virtually from the images.

What our technician does

1. Checks the state of the array and each drive, and the alert settings.
2. Assesses what data is stored and how long ago backups were made.
3. Sets up backups with version history to a separate device and to the cloud.
4. Turns on monitoring and drive health alerts for the responsible staff.
5. After a failure, recovers the array from drive images without risking the originals.

A real example

A small logistics company in Navtlughi kept all its documents and its 1C database on a NAS with a two-drive mirror and believed it didn't need a backup. One drive had failed more than a year earlier, but the alerts were going to the inbox of an employee who had left. When the second drive started failing, the NAS stopped opening. We managed to recover most of the data from the damaged drive in the lab. Afterwards we fitted the NAS with new drives from different batches, set up alerts to the director and the accountant, daily backups with version history to a separate drive, and weekly backups to the cloud.

Remember too that RAID doesn't speed up recovery after a serious disaster. If a server is hit by flooding or ransomware, only a backup can get work going again, and it's important to know in advance how long restoring from it will take. For a business, it's worth running a practice restore once: bring the backup up on a separate computer and make sure the documents and database open.

Call a technician

If you have a RAID array without a backup, or the array has already failed, we'll help in any district of Tbilisi around the clock. Learn more on our data recovery, hard drive recovery and local network setup pages.