A degraded array limps on until another failure occurs (for example, a second disk fails), a rebuild goes bad, or the controller dies along with your RAID configuration. We restore RAID 0, 1, 5, 6 and 10 arrays from servers, workstations and NAS units throughout west London and Surrey, and we do urgent RAID server recovery for businesses.
A free diagnostic for every RAID array job. Once we have completed our diagnostic, one fixed price quote is put in writing before any work starts.
In most cases you only pay if we recover your data. For all electronic and mechanical work, including chip-level repairs, 50% is required upfront, with the balance only payable if we recover your data. For BitLocker, ransomware, CCTV/DVR and forensic jobs, full payment is made upfront once you accept the fixed quote. Full details regarding our costs can be found on the data recovery cost page.
To start any RAID job, we match what is failing to identify where the fault is. There are twelve common issues that account for almost everything that reaches our bench.
The worst thing about RAID: it was designed so that a single loss is survivable, but the rebuild that follows then finishes off a second, older drive.
If you run a rebuild onto a suspect drive, it may stall partway through and damage the very parity the rebuild is depending upon.
The card (controller) takes all of the array configuration with it when it dies, so the geometry needs to be worked out from the individual drives.
If you force an outdated drive back into an active RAID array, the file system will shred itself due to the inconsistencies between the old and new data.
Do not let a new RAID array be initialised over the existing one, as it will destroy the RAID metadata. Stop it before initialisation is completed.
A power event can cause two or three drives to fail simultaneously; each one will need to be individually repaired and imaged before anything else happens.
If you find that multiple drives "failed" at nearly the same time, check the chassis wiring before the disks: diagnose first, panic never.
The RAID itself is fine, however the file system (NTFS / ext4 / VMFS) above it is damaged, and this is rebuilt at parser level.
As long as the space underneath a deleted definition remains intact, the definition can be recovered, right up until those sectors have been written over with new information.
If your controller battery is flat and a sudden loss of power knocks parity and data out of sync, this is resolved at image level.
When a Dell PERC or HP Smart Array controller fails, it loses the RAID definition, so we derive the geometry using only the RAID group members themselves.
When one member's heads fail and it starts clicking, we repair it on the head-swap bench and image it like the rest of the set.
Every RAID group member is captured in a total bit-for-bit copy before a single decision is made about how to repair it, and we never touch the original member. We use each member's image to determine the geometry (order of members, block size, rotation and offset). From there we create a virtual volume based upon the geometry and extract the file system sitting on top. Nested RAID 10s, RAID 5s carrying a stale member, RAID 6 double failures, hardware controllers that forgot their configuration: these are all routine jobs in our lab.
Do not force drives back online. Do not run a rebuild onto suspect disk(s). And don't let an overzealous generic "IT fix-it" try "just one more thing": partial rebuilds destroy the parity recovery depends upon. Turn it off. Mark each drive with its bay number. Call us. If your RAID is down and it's business-critical, tell us when we answer the phone and we'll expedite the diagnostic process.
There's only one rule for every RAID job (nothing happens until everything is imaged), and our bench exists to enforce this rule:
Every drive is duplicated at the same time onto its own independent piece of equipment, before anyone analyses any information on the individual drives; no original drive ever goes back into any kind of rebuild process.
Analyses each copy to determine how the drives were configured as part of the RAID array: which bay each drive occupied, the block size, the rotation and the offsets. Once those parameters have been determined, it brings up a virtual version of the RAID array.
Failed members don't get left behind either: they get the complete hard drive treatment (heads, firmware and electronics) until they're able to be imaged along with the other members of the RAID.
Once the above has been completed, the resulting geometry is verified via parity and entropy tests before the files are restored from the reassembled volume.
Whatever lived on top of the array (NTFS, ReFS, ext4, XFS, BTRFS, VMFS, even virtual machine disks) is read natively.
The members of the array are never written to: they're only ever read and go home in the state they arrived.
Whether it's hardware, software or somewhere in between, the geometry lives on the members themselves and that's all we need. A failed Dell PERC or HP Smart Array controller loses the definition of the array, which we derive from the members as well; and where a Seagate or WD member has failed heads or firmware, it is repaired on the head-swap bench and then imaged before the array is rebuilt.
Turn off the server or array, remove the member disks and label each one with its bay or slot number. Without that order, rebuilding the system gets much harder. Call us on 0800 689 0668 or use the free online diagnostic for a quote before posting. Only the labelled disks need to be sent, so you can keep the controller and chassis where they are. A quick photo of how the disks were originally laid out is genuinely useful too.
Before sending anything, get a quote from us either by telephone or using our free online diagnostic. Then post your hardware to us (tracked and insured), or drop it off at our Guildford laboratory in person.
Sending a drive from a computer, laptop, MacBook, iMac, CCTV/DVR or server? You will need to remove the internal hard drive or SSD and send us just the drive, as we are unable to remove the drive for you. We also cannot recover storage soldered onto a motherboard, such as in Apple Silicon Macs and some thin laptops. Only drives that can be removed from a machine and sent to us can be recovered. If you are unsure whether your device has a removable drive, please call us before sending. RAID, server and SAN drives need to be removed from their enclosure and labelled by bay. For a NAS, send the disks themselves if they are removable, otherwise the entire NAS device. An external single drive should come to us whole, along with its cable and power supply.
↓ Print the booking-in & shipping form (PDF)
Please mark the package for the attention of Guildford Data Recovery, and we will call you as soon as we have diagnosed your media.
Unsure as to how to proceed with sending the item in for recovery? Then simply call 0800 689 0668, or complete the free online diagnostic prior to dispatching the item to obtain your price quotation.
Free diagnostic, fixed quote, and in most cases you only pay if we recover your data. Start now or call the freephone.