RAID levels · explained

RAID 10, explained properly.

RAID 10 is the setup a lot of businesses reach for when they want both speed and safety — and it delivers both, which is why it’s so common in servers and NAS boxes. But “safe” gets misunderstood, and RAID 10 arrays do fail and lose data. Here’s what RAID 10 actually is, what it protects against, and — importantly — what it doesn’t.

Free 48-hour diagnostic
Handled in-house
No fix, no fee · most jobs
// in short

Fast and redundant.

RAID 10 combines mirroring and striping — quick and resilient. But it tolerates only the right disk failures, and it’s still not a backup.

1+0
Mirror then stripe
4+
Minimum disks
Survives
One disk per pair
Not
A backup
// what it is

What RAID 10 is.

RAID 10 — also written RAID 1+0 — is a combination of two simpler ideas. Mirroring (RAID 1) keeps two identical copies of data on a pair of disks, for safety. Striping (RAID 0) spreads data across multiple disks, for speed. RAID 10 does both: disks are grouped into mirrored pairs, and data is then striped across those pairs.

It needs an even number of disks, at least four. The result is an array that reads and writes quickly (thanks to striping) while keeping a full second copy of everything (thanks to mirroring) — the best of both, at the cost of using half your total capacity for the mirror.

// how it works

How it works.

Picture four disks as two pairs. Your data is split into stripes; each stripe is written to one pair, and within that pair it’s written to both disks at once, so there are always two copies. Reads can come from either disk in a pair, which makes them fast; writes go to both.

Because the data is duplicated within each pair and shared across pairs, RAID 10 gets its speed from spreading the load and its safety from never having just one copy of anything — provided the failures fall the right way, which is the crucial detail.

// redundancy

What it protects against.

This is where RAID 10 is often misunderstood. It can survive one disk failing in each mirrored pair — potentially up to half the disks in the array — because the other disk in each pair still holds a full copy. That’s genuinely resilient.

But there’s a catch: if both disks in the same pair fail, that pair has no surviving copy, and the whole array is lost — even if every other disk is fine. So RAID 10 doesn’t simply “survive half the disks failing”; it survives the right half. Two failures in one pair, and it’s gone.

// how it fails

How RAID 10 fails.

Even a resilient array fails, and we recover them regularly. The common causes: two disks failing in the same pair; a failed rebuild after replacing a dead disk, where a second disk gives out under the strain; controller failure that scrambles the array configuration; human error like pulling the wrong disk; and file-system corruption on top of the array.

Disks bought together tend to age together, too, so multiple failures in a short window are more common than people expect — which is exactly how a “safe” array ends up needing recovery.

// not backup

Why it’s not a backup.

The most important point of all: RAID 10 is redundancy, not backup. It protects against disk failure, but it does nothing against the things that actually cause most data loss — accidental deletion, ransomware, file corruption, a controller failure, fire or theft. Delete a file on a RAID 10 array and it’s deleted on both mirror copies instantly.

So RAID 10 keeps a server running through a disk failure, which is valuable — but it’s no substitute for a proper, separate backup. Treat it as uptime insurance, not as your only copy.

// recovery

How RAID 10 is recovered.

When an array does fail, recovery doesn’t mean rebuilding on the live disks — that’s risky. Each disk is imaged individually to healthy storage, the array’s structure (which disks pair up, the stripe order and parameters) is reconstructed virtually from those images, and the data is extracted from the rebuilt array — all without stressing the original disks.

Done this way, most RAID 10 failures are recoverable, even multiple-disk ones — provided the array is powered down promptly rather than repeatedly rebuilt.

// faq

Common questions.

What people ask us most about RAID 10.

It depends which disks. RAID 10 can survive one disk failing in each mirrored pair — up to half the array — because the other disk in each pair holds a full copy. But if both disks in the same pair fail, that pair has no surviving copy and the whole array is lost. So it tolerates the right failures, not simply any half of the disks.

No — and this is the most important thing to understand. RAID 10 protects against disk failure, but not against deletion, ransomware, corruption, controller failure, fire or theft, all of which are common causes of data loss. Anything deleted or corrupted is affected on both mirror copies at once. RAID 10 keeps a system running through a disk failure, but you still need a separate backup.

Usually yes. Whether it’s two disks failing in a pair, a failed rebuild, or a controller fault, recovery works by imaging each disk, reconstructing the array’s structure virtually from the copies, and extracting the data — without rebuilding on the fragile originals. Most RAID 10 failures are recoverable, provided the array is powered down promptly rather than repeatedly rebuilt.

// RAID 10 failed?

RAID 10 array down? Don’t rebuild it.

Power it down and don’t keep rebuilding — that stresses fragile disks. Send it in and we’ll image the disks and reconstruct the array safely. Post it in from anywhere in the UK, or drop it to us in Belfast.

Call us — 028 9002 0144
Mon–Fri · 9am–5:30pm · No fix, no fee
Start a free diagnostic →
028 9002 0144