The bench is open for new cases · Mon–Fri, 9am–5:30pm Need it fast? Ring 0800 6890668
EDR Exeter Data Recovery 0800 6890668 Start my case
EDR / Specialist / Recovering data from servers

Specialist · server data recovery

Server data recovery in Exeter. You can order a new server today, not a new copy of what it held.

Most failed servers still have functioning hard drives. What is broken is the state surrounding them: a RAID controller in conflict with its own metadata, a prompt that will finish off your RAID array as soon as you enter a single incorrect answer, or data writes which were cached but never made it to disk. We rebuild the volumes, databases and virtual machines that an Exeter company actually runs on, using images of the disks, without ever needing the original controller.

Only pay if we recover your data, on most jobs Free diagnosis and a written quote Get a quote first, then post from anywhere in Devon

Talk it over with an engineer
0800 6890668

What the server's symptoms are telling you.

Not there? Head for the triage →
What's wrongWhat it usually meansStep one
Dell PERC: Foreign Configuration DetectedDisk metadata and the card's memory are out of sync; the controller has stopped for safety reasons and not due to lost data.Decline both prompts.
1786: Drive Array Recovery Needed (HP)Either a queued rebuild or one which was in progress when it failed.Shut down power; image each disk before proceeding.
A power cut, and now the bays are glowing amber.Writes were "in flight" at the time and the cache became questionable; the layout of the array is still on the disks.Do not force anything back online.
Boots OK, however the data volume is missing.Either damage exists in a controller state or there is file-system corruption; the data "below" it is still present.High chances that your data will be recovered
The controller card itself has failed.All stripe maps, disk orders, and parity-rotation information are stored on each of the individual disk members.You don't need another identical card.
Files have been recovered but they won't mount (VMs, databases)Recovered does not mean consistent; consistency is the true testSee also our page on VMware, and our database page
Posting it to us: get a quote first, by phone or with the form, then please use a traceable (tracked) and insured postal delivery method to send your device to our intake lab in Bristol; the return is paid for by us. Alternatively you can drop your item off there yourself, Monday to Friday, 9am to 5:30pm, with no appointment needed. An engineer will talk you through how to pack your unit when you call to obtain the quote. More information about sending your items can be found on the contact page.

The map is on the disks, not the card.

Dell PERC & MegaRAIDBoth write metadata in DDF, a standard for the industry, into the final sectors of each disk. The record is used by the RAID to be able to rebuild itself from images, without any cards being attached.
HP’s Smart ArrayHP chose their own path: proprietary RIS metadata at the beginning of every disk, and parity data offset by one stripe. Any generic recovery settings fed this input will produce garbage.
Import or ClearIf you time this incorrectly, either option will be destructive. If you choose Import, it may write stale metadata onto live stripes. If you choose Clear, then the striping map will be gone from the drives forever. Once we have taken an image of each drive, both options lose their teeth.
Cache that never landedWhen a controller accepts write commands but they are still sitting in the battery-backed RAM and fail during that flush of data, the corruption resides deep in the filesystem; hence “it simply lost power” is a candidate for imaging, never a hopeful repair.

The recovery, one stage at a time.

01

Booked in, assessed for nothing Free

Your device will be logged under its own case number on the day it comes into our lab. The engineer who works on your device identifies the problem with it, clearly states what we think has a good chance of coming back and provides you with one fixed quote. There are no charges for this assessment; there is no obligation for you to do anything; and paid work does not begin until you authorise us to start.

The look is freeA single quote, in writingNo obligation
02

Freeze the scene

Turn power off. Label each disk by its bay number. Do not answer any prompts. Each member is then independently imaged using specialised imaging equipment (including those the controller deemed as 'dead') before one byte is interpreted.

Every member imagedFailed members handled
03

Put the array back together in software

The images will have the DDF or RIS records in them. The DDF or RIS record contains the RAID level, striping unit size, physical disk number (disk order) and, if you are running RAID 5 or 6, the parity units with their rotation information. From those records the array is rebuilt virtually. Losing the original card is simply irrelevant as far as rebuilding the array is concerned.

Geometry taken from the membersWorks without the old card
04

Prove it works — not merely that it copies

On the rebuilt virtual copy, we repair everything which was stored on top of it. If that is an NTFS/ReFS filesystem, a VMFS datastore or an ext4/XFS filesystem, then once that is repaired we verify whether your "guests" boot properly, whether your databases attach correctly, and whether your shared resources mount properly.

Repairs run on the copiesBoots and mounts, demonstrated
05

Watched back, verified, returned

You have access to a complete list of every item recovered from your device, which you approve prior to any fees being applied. The files will be returned via new media (brand new) with the return postage paid by us, and the case remains open for as long as it takes to confirm that all the files open properly on your own machine.

You approve the file listReturned on new mediaWe pay return postage

Opening checks at the bench

  • RIS opens a disk, DDF closes it: HP starts at the front end of the disks while PERC and MegaRAID start at the back end; knowing this determines whether we are in rebuild or guess mode.
  • A ‘failed’ disk yields useful information for the rebuild: a disk that dropped weeks ago may have the most up-to-date version of certain blocks, so each block will be weighed against all of its peers during the rebuild process.
  • Delayed parity on an HP system confuses generic rebuild code: Smart Array delays parity by one stripe, so standard rebuild presets produce garbage. The rebuild must use the same parity structure as HP’s own.
  • Ask early about the cache battery: the cache battery is an important part of the process, because if the write-back data made it to the platters, the file system will be cleaner. Therefore, before making commitments to anybody, we determine this.

Import vs. Clear (what they really do): Import imports the foreign data's metadata to overwrite your current running config. Clear clears out all members' DDF maps for good. Both are one keypress; there is no "undo". Once you have an image of each disk, Import and Clear cannot harm you. This is why we start with imaging: this is not precautionary, it is the entire strategy.

Fresh from the casebook.

EX · EDR-2026-2816VERIFIED ✓

A storm spared this ProLiant — the restart didn't

The power outage was longer than the UPS allowed for. So, when the server came back on line, the Smart Array was asking questions that no one could safely respond to. Therefore, we made images of each of the four disks, extracted the RIS information and honoured the delay in parity; finally we mounted the practice-management database from the rebuild, with the data intact up until the point that the last write in cache was flushed just prior to the loss of power.

Database mounted clean5 days in the lab

Before it leaves your hands.

Do

  • Take photographs of the controller screens, and then shut it down.
  • Document the bay number on each disk as you remove it.
  • Always get a quote first, then send all the disks in the set; even the ones that look like they are dead will still have something to contribute
  • Identify what was being served by it: SQL, Exchange server, guests, networked shares

Steer clear

  • Clicking “Import” or “Clear” when that prompt appears.
  • Placing a spare disk so you can begin a rebuild.
  • Doing firmware updates on the card during an emergency.
  • Starting the server again when there is no image available.

Straight answers from the bench.

The server won't boot — has the data gone?

Generally no. Most times it's because of controller state issues, array metadata corruption or file system corruption. All these are located on disks that have no physical damage. Shut off power to the unit. Do not click through any prompts. Your chances will remain very high.

'Foreign Configuration Detected' — what's it telling me?

Your Dell PERC is being cautious because it has read disk metadata which doesn't match up to what its memory thinks it knows. This is typical if you have switched out some of your boards, changed the order in which your disks are installed, or there was a controller failure. Stopping is a deliberate act of caution on its part. With both options on the screen, destroying data is possible, so do create an image of all your disks before deciding what to do next: once images exist for all the disks involved, neither option will damage anything.

Can you rebuild the array without the original controller?

Yes. The RAID layout lives on each disk itself. While PERC and MegaRAID use a standard DDF format to store their RAID layouts on disk, HP uses a different format called RIS. Because we are rebuilding on the bench from images taken from the original disks, we don't need an identical controller.

Do VMs and databases come back, or just plain files?

Both, and there is a difference worth making. Being able to get the virtual machine files (VMDK) and database files (MDF) back is about half of the work; getting those virtual machines and files mounted by the hypervisor or database engine is the second half. It's exactly the layer that our VMware and database pages describe.

More like this in the casebook.

RAID 5 arrays → Recovering RAID 0 → Database recovery → What we take in →

Whatever the fault, leave the power off.

Every time a failing hardware component is powered back up, there is additional lost data. First open a case with us: the inspection (look over) is always at no cost, whichever way it goes.

0800 6890668

Get a free diagnosis Sending a device in