🇧🇪 Dafotec Belgium · 3 drop-off points · ISO 5 lab in Roubaix
DAFOTECsince 2004 Estimate my price
Service · RAID & server

Data recovery from a RAID server

Degraded array, failed rebuild, dead controller, lost configuration. When a server goes down, every hour costs. We virtually rebuild your array — RAID 0, 1, 5, 6, 10, 50, 60 — without ever writing to the original disks, with a dedicated enterprise emergency desk.

Flat feefrom €700 excl./disk
RAID 5 — 1 disk down95–98%
Enterprise emergency24–48h
Diagnosisfree · 24h
Symptoms we treat

Does your server show one of these signs?

Multiple disks down

Beyond the RAID level's tolerance, the array goes offline. Often, not every disk is actually dead.

60–75% with 2+ disks affected
Rebuild interrupted / failed

A rebuild launched on a failing disk can overwrite the other members. Stop everything immediately.

array often recoverable
Lost configuration ("Foreign")

After a controller reset or board swap, disks show as "Foreign" and the array refuses to mount.

95–98% if platters intact
VMware / Hyper-V volume lost

Corrupt or missing virtual machines (VMDK, VHDX, QCOW2) on top of a degraded RAID.

VM extraction after rebuild
Our method

What the laboratory actually does.

Absolute principle

We never touch the original disks

Each member disk is cloned sector by sector before any analysis, using parallel imagers (8+ ports) and write blockers. All reconstruction then happens on the copies — the physical array stays intact for any later attempts.

RAID reconstruction — de-striping

Hex analysis of the array

Without trusting the controller, we determine the array's real parameters by analyzing the entropy of the data: stripe size, exact disk order, parity rotation scheme and initial offset. RAID 5/6 parity is recomputed via an XOR operation, letting us virtually reconstruct the missing block of any disk.

Underlying disks

Repair of failed members

Some disks need work before they can be read: head transplant in an ISO 5 cleanroom, PCB swap with ROM transfer, Service Area access. We hold over 20,000 spare parts in stock to speed the operation.

Virtualization

Virtual-machine extraction

Once the volume is rebuilt, we mount the file systems (VMFS, ZFS, ReFS) and extract the virtual machines — VMDK for VMware, VHDX for Hyper-V, QCOW2 for Proxmox — along with SQL databases (Server, Oracle, MySQL, PostgreSQL).

Business continuity

Enterprise emergency 24-48h

Diagnosis and recovery run around the clock, a dedicated contact from diagnosis to return, NDA on request, processing on an air-gapped network. Return on a new encrypted device.

Proof in the numbers

RAID success rates across 120,000+ cases.

Averages observed since 2004. RAID is about 30% of our enterprise activity.

RAID 5 — 1 disk down98%
RAID 6 — 2 disks down95%
Lost config / dead controller95%
RAID 0 — 1 disk failed72%
RAID 5 — 2+ disks down68%
Coverage

All controllers, all levels.

Dell PowerEdge / PERCHP ProLiant / Smart ArrayIBM ServeRAIDLSI MegaRAIDAdaptecSupermicroLenovoFujitsuArecaIntel RST

RAID 0, 1, 5, 6, 10, 50, 60 levels, hardware and software. VMware ESXi, Hyper-V, Proxmox virtualization. VMFS, ZFS, ReFS, EXT4, XFS, NTFS, Btrfs file systems.

What you must never do to a failing RAID server

  • Launch an automatic rebuild — a rebuild on an unstable disk massively overwrites the other members.
  • Swap the disk order around — order is critical array data; losing it badly complicates reconstruction.
  • Reset the controller — the configuration (stripe, parity, offset) can be erased.
  • Replace a disk "to see" — any write to the array lowers the odds.
  • Force the volume to remount — a file system mounted read-write on an inconsistent array worsens the corruption.

The golden rule: power the server off, note the bay order if possible, and send the whole unit — disks and controller — to the lab.

FAQ

Specialist answers on RAID.

How do you recover data from a RAID 5?+
RAID 5 spreads data and parity across all disks and tolerates one disk failure. Beyond that, or if the controller fails, the volume becomes inaccessible. We clone each member sector by sector, virtually rebuild the XOR parity and extract the data without writing to the originals. Standard turnaround 48 to 72h.
The rebuild failed — how do I limit the damage?+
Shut the system down immediately. An incorrect rebuild can overwrite RAID metadata and the data of other members. Don't relaunch any rebuild before diagnosis — advanced techniques can often restore the original configuration.
Is a disk shown as "Foreign" lost?+
No, not necessarily. Very often, only the configuration metadata is at fault, not the data. As long as nothing has been rewritten, the array remains rebuildable.
Do you recover virtual machines?+
Yes. After rebuilding the volume, we extract VMDK (VMware), VHDX (Hyper-V) and QCOW2 (Proxmox) VMs, along with SQL databases, even on corrupt file systems (VMFS, ZFS, ReFS).
How much does RAID recovery cost?+
The flat fee starts at €700 excl. VAT per disk handled (up to €950 excl. VAT/disk for complex cases). Free diagnosis, optional 24-48h emergency, no data – no fee: only €25 due if we fail.
Every hour of downtime costs

RAID server down? Act right.

Power the server off, launch no rebuild, don't swap the disks. Request priority handling — free diagnosis, enterprise emergency 24-48h.