Open, and work is coming in · weekdays, 9am–5:30pm Faster by phone: 0800 6890668
PDR Plymouth Data Recovery 0800 6890668 Get a price
PDR / The jobs we take in / Servers and RAID arrays

Servers · stopped RAID sets

Server and RAID recovery, Plymouth. The chassis has stopped; what is inside it is usually in good order.

The drives are normally the sound part of it. What has failed is everything around them — a card whose record of the array no longer agrees with the disks in front of it, a prompt where both buttons cost you something, a cache holding writes the mains interrupted. Each member is copied first and every repair after that happens on the copies. Leave your controller where it is. What comes home is the volume entire: file shares, mail stores, databases, virtual machines. A fish merchant's office at Coxside and a research spin-out at Plymouth Science Park send in the same chassis.

Nothing readable? Most jobs carry no bill Free diagnosis, then one written price Boxes come to the lab from Truro, Newquay and Tavistock

You will be talking to an engineer
0800 6890668

Server symptoms, and what each points to.

Yours not here? Start at the triage →
What shows upWhat that points toWhat to do
PERC prompt: Foreign Configuration DetectedWhat the disks say about the array and what the card remembers no longer match, so it halts instead of guessing. Nothing has goneGive it no answer at all
1786 — Drive Array Recovery Needed (HP)A rebuild began and might have stopped midwayPower it off. Copies before anything
Amber bay lights following a mains failureWrites were cut off half-done and what the cache still held is anyone's guess. The array's own description is intactDo not switch it on again
The machine boots; the data volume is missingThe fault sits in the controller's configuration or in the file system; what lies beneath both is untouchedUsually recoverable
The RAID card has goneBay order, stripe size and parity rotation are all written on the membersNo matching card required
Files copied off, yet no VM or database will runCopied is not the same thing as consistent, and only consistent data will startThat becomes a database recovery job
Getting it to us: wrap it so nothing can shift, cover it for what it is worth, and post it tracked to the intake lab in Bristol. We cover the postage back. If you want an engineer to check the packing before the box is taped, ring first. It is spelled out on the contact page.

The disks carry the array's own map.

MegaRAID & PERC cardsDDF is an open format and both these cards write it into the closing sectors of every member. Lift it out of the images and the array reassembles with your controller still sitting in the server.
HP's Smart Array cardsHP goes its own way. RIS structures sit at the start of every member, and its parity block lands one stripe off the customary position. Point a default preset at that and out comes a volume your server never had.
The prompt with two bad answersLeave this one alone. Import is able to write stale metadata across stripes that were perfectly sound, and Clear lifts the description off every disk with nothing to restore it from. Copy the members first and the prompt loses its teeth.
Writes stuck in the cacheA write is reported complete the moment battery-backed memory holds it, which is a long way before any platter sees it. Lose the mains mid-flush and the harm settles inside the file system where nothing displays it. That alone is the case for copying first.

What happens to a drive while it is here.

Cases in the log →
01

A number on arrival, then the free diagnosis Free

Whatever arrives is booked in under a case number on the day it reaches the bench. An engineer then works out the actual fault, and that part is free. You get a plain answer on what can come off the drive and what cannot, followed by one price in writing. Nothing further happens until you have read it and agreed.

Diagnosis at no chargeA single written figureYou owe nothing yet
02

Leave the machine as you found it

Shut the server down. Take the drives out one at a time, writing each bay number on the drive that came from it. Whatever the console asks, give it nothing. On the bench every member is copied separately, on hardware made for the job — the ones the card threw out included. No file gets opened until the last copy has finished.

Every disk copied separatelyThe dead disks as well
03

Reassemble the array in software

DDF or RIS records travel across inside the images. Read them and out comes the RAID level, the size of a stripe, which bay held which member and, on any parity set, the direction parity travels. That is the whole geometry, and enough to build the array in software. Once it is built, a failed card stops being anybody's problem.

Geometry taken from the membersYour card stays in the server
04

Copying it is not the same as opening it.

Whatever was layered above the array — XFS, ext4, NTFS, ReFS, a VMFS datastore — gets its repairs done on the reassembled copy, never on the drives you sent. Then it is tested in earnest: guests boot, shares mount, databases attach, and individual files are opened.

Repairs made on copies onlyProved by mounting and booting
05

You give the word, and it ships back

Thinking it over costs you nothing. Everything the drive gave up is listed for you before any invoice exists, and the bill only follows your go-ahead. Files travel back on media bought in for your job, with return carriage paid at this end, and the case stays on the bench until you confirm they open on your own computer.

You see the list and decideWritten to fresh mediaWe pay the postage home

The faults we see most

  • Not every card writes its map in the same place — HP puts RIS at the front of each member, while PERC and MegaRAID keep DDF in the closing sectors. Read the wrong end of a disk and the geometry becomes guesswork.
  • A discarded drive still counts — the member a card threw out weeks ago can hold the only readable version of some regions, so the images are set against each other byte by byte. Nothing on this bench is accepted on trust.
  • Default presets get Smart Array wrong — HP sets the parity block one stripe away from where the usual tools look for it, so the volume such a tool produces never existed on your server. It has to be built to HP's own scheme.
  • The cache battery comes up early — whether anything held in write-back ever reached a platter settles which state the file system is in, and that is answered before anyone is handed an expectation to hold on to.

What those two buttons actually do: Import accepts the foreign description as correct and stamps that across whatever configuration the card is running today. Clear strips the description off every member, with nothing to put it back. Both are one keypress away and the screen gives you no warning at all — though once full copies exist, neither can cost you anything. Copying is not one stage among several on this kind of job. Everything else stands on it.

A job out of the casebook.

PL · PLY-2026-0816LOGGED ✓

A dead cache battery, and the practice database intact

The mains had been down long enough to drain the cache battery flat. On restart the controller asked the practice a question none of the staff could safely answer, so it sat there unanswered. Nothing was run on the disks themselves: every member was imaged first, and the array's own metadata came back off those images. The parity rotation followed from that, and the set was assembled on the copies. The practice database opened whole, out to the final write the cache had held.

Database read back whole5 days start to finish

Before you seal the box.

Get these done

  • Photograph each screen the card shows you
  • Number each drive to the bay it occupied
  • Send every disk, the apparently finished ones as well
  • Tell us the workload — SQL, Exchange, guests, shares

What to avoid

  • Choosing Import or Clear at that prompt
  • Dropping a spare in and kicking off a rebuild
  • Putting fresh firmware on the card at this point
  • Switching the server on before copies are taken

Questions that come up most weeks.

The server will not start. Have we lost everything?

Seldom. The disks themselves are generally sound and the fault sits a layer up: the card's own record of the set, the description of the array, or a file system that has been chewed. Keep the power off, answer nothing the console asks, and the good position you are in stays that way.

What is 'Foreign Configuration Detected' telling me?

That what is written on the disks differs from what the card expected to find. A replacement controller, a disk returned to the wrong slot, or a card going bad will each produce it. Halting there is the card behaving properly. Both Import and Clear destroy things, so copy every member before you decide anything. Once copies exist, neither choice can take your files.

Must the controller travel with the disks?

No. Each member carries its own account of the array. DDF, a published format, is what PERC and MegaRAID write; HP prefers RIS structures of its own. Whichever it is gets read out of the images, and the array is put together in software. No card is needed for that.

Will the databases and the virtual machines start again?

Generally, though it is two pieces of work, not one. First the VMDK or the MDF has to come off the reassembled volume whole. Then a hypervisor has to be talked into mounting the one, or SQL Server into attaching the other. The second half is what our database recovery page covers.

Pages people open after this.

RAID 5 rebuilds → RAID 0 sets → MacBook Air storage → All the media we take →

A drive that stays switched off gets no worse.

The diagnosis is free, and it comes back as a list: which files read, which do not, and what getting them off would take. Until then, leave the drive unplugged.

0800 6890668

Book a free look Sending it in to us