Open, and the first look is free · 9am–5:30pm, Mon to Fri Quicker to ring: 0800 6890668
RDR Reading Data Recovery 0800 6890668 Get it quoted
RDR / What we take in / Dell PowerEdge servers

Specialist · Dell PowerEdge and PERC arrays

Dell PowerEdge recovery, Reading. Cut the power before the halted rebuild asks more of the next-weakest disk.

Servers and RAID sets start at £500 + VAT, the diagnosis is free, and you have the figure in writing, fixed, before any chargeable work starts. What arrives is nearly always the same shape: a PowerEdge that has quietly been a member down for weeks, a spare fitted, a PERC rebuild allowed to run, and a second carrier going amber part-way through it. Every disk is then cloned read-only on its own, the controller's own account of the set is read back off those images, the array is assembled above them and the file system repaired on the copy. Your carriers are never written to, and the chassis stays where it is. Sets come in from Reading itself, and from Newbury, Thatcham, Theale, Woodley and Didcot.

Nothing recovered? On most jobs, nothing to pay The diagnosis is free; then one fixed figure in writing Parcels arrive from Newbury, Woodley and Theale

An engineer answers, and looking is free
0800 6890668

Match the PowerEdge symptom to the fault.

Different fault? Begin at the triage →
What you are seeingWhat is likely wrongWhat happens next
A rebuild started and stopped part-wayA second member began to fail while the set was already short of onePower the chassis down now
The PERC reports a foreign configurationThe controller has lost its own description of the array after a power eventImport nothing, and ring first
Two carriers went amber at the same momentMore often a backplane or expander fault than two disks failing togetherTake it off the mains, send every disk
The virtual disk shows as offlineA member has been marked failed and the set has stopped answeringDo not clear the configuration
The disks read as nothing on a PCA SAS drive needs a SAS port a desktop does not have, and a PC sees the PERC's own metadata on each member rather than a filesystemNormal, not damage. Send them anyway
Writes were still in cache when the power failedA flat battery on the cache module loses whatever it was holdingSay so — it changes the order of work
Packing and posting: pad it so nothing can shift, put a value on the parcel that reflects what the data is worth to you, and send it tracked to the intake lab. Return postage is ours. Ring first if you would rather an engineer talked the packing through before you seal the box. It is all set out on the packing and postage page.

Where a PowerEdge array actually fails.

PERC controllers, H310 to H755Dell's own RAID cards, and the part that decides most of these jobs. A PERC writes its account of the array — member order, stripe size, parity rotation, where the data begins — onto the disks themselves, which is why a set can be put back together here from images without the card, the backplane or the chassis it came out of.
SAS, SATA and 520-byte disksA PowerEdge chassis mixes drive types freely. A SAS member needs a SAS port a desktop does not have, and even on the right port a PC finds a RAID member carrying the PERC's own metadata, not a filesystem it can mount. That is the controller at work rather than damage. Sectors of 520 bytes, where they turn up at all, are a storage-array habit — NetApp, EMC, 3PAR — rather than a PERC one, and both kinds are ordinary work on the right equipment.
Boot mirrors and hot sparesMost of these servers start from a mirrored pair and keep the data on a separate RAID 5 or 6, often with a spare sitting idle in a bay. The mirror seldom causes trouble; the spare frequently does, because the moment a member is marked failed the controller begins writing to it and the array is busiest at its weakest hour.
What we never do hereNo rebuilds, no imports, no clearing a configuration, no forcing a member back online against the drives you posted. Each of those is a write, and a write is what closes options. Your members get read once and put aside, and every repair happens on the copies, where a wrong turn costs an afternoon and nothing else.

From your parcel to your files coming back.

Jobs we have closed →
01

Logged the day it lands, and the look costs nothing Free

A case number goes on your device the day it reaches us, and an engineer then finds the real fault — free, and before anything else happens. You are told plainly what will come back and what will not, with one fixed figure in writing beside it. Nothing chargeable starts until you have read that and said yes.

The look costs you nothingA fixed price, in writingStill nothing to pay
02

The disks are copied before they are asked anything

Each member is read behind a write blocker and cloned in full before the array is so much as looked at — the sound ones as well as the one the controller wrote off. Awkward regions get brief runs and wait until the rest is safely down. Not a byte goes back onto the drives you posted, at any stage of the job.

Every member cloned firstNo write reaches your disks
03

The controller's own account, off the copies

A PERC writes its description of the set onto the members themselves, which is why neither the card nor the chassis is needed here. Member order, stripe size, parity rotation and the offset the data starts at are all read back from the images, then checked against the file system sitting above them before anything is assembled.

Set assembled on the imagesBrought up read-only
04

The volume, and then the file system

With the geometry settled, the array is put together over the copies and brought up read-only, and the file system is repaired there rather than on anything of yours. Files are then opened — databases, shares, mail stores, accounts — because a name in a directory listing proves nothing at all. Whatever comes back damaged is listed as damaged.

Repaired on copies onlyOpened here before it goes
05

The list first, then the bill, then the files

You see the file list first, then decide. Nothing is invoiced until you have said yes, and on most jobs there is no fee at all if the data does not come back. What we pull off goes onto media bought in for your job, and the postage home is ours. The case stays open on the bench until you tell us the files open on your own machine.

Nothing billed until you say yesNew media, bought for the jobPostage home is on us

What goes wrong most often

  • A rebuild is the most demanding thing you can start — it reads every sector of every remaining member, and those members are the same age, from the same batch, with the same hours on them. A disk failing during a rebuild is not bad luck; it is the second-weakest disk being asked to do the hardest work it has ever done.
  • A server that will not start is often a healthy array — a PERC, a backplane, an expander or a power supply can take a whole chassis out of service while every disk inside it is sound. That is a cheaper job than it looks from where you are standing, and the free look settles which of the two you have before a figure is agreed.
  • Disks reading as nothing on a PC are normal, not damage — a SAS drive needs a SAS port a desktop machine does not have, and even on the right port a PC finds a RAID member carrying the PERC's own metadata, not a filesystem. A good many people conclude the disks are finished on exactly that basis. They are not, and they read here on their own terms.
  • Send every disk, the written-off ones included — a member the controller dropped is usually far more readable than the controller's verdict suggests, and it often holds the newest copy of the blocks that matter. Bay numbers on each drive, and an honest history of what has been tried, decide how quickly the set goes back together.

What decides these jobs: how much of each member still reads, whether anybody re-created the virtual disk after it failed, how far a rebuild was allowed to run, and whether the volume was encrypted. A chassis powered down with all its members still in it gives the longest list of possibilities. That same chassis after a spare, a rebuild, an import and two nights of restarts gives a much shorter one — and you are told which files fall in the gaps rather than finding out at home. An encrypted volume needs its key from you. Storage is all we handle: no phones, no tablets. Servers and RAID sets are from £500 + VAT, quoted in writing after a free diagnosis, and on most jobs there is nothing to pay unless the data is actually back.

How one of these actually went.

RG · RDG-2026-2007ON FILE ✓

A rebuild that halted on a six-bay PowerEdge, and two volumes back

An engineering consultancy at Didcot had run a six-bay PowerEdge a member short since the summer without anybody noticing. When a second carrier went amber, a spare was fitted and the rebuild allowed to run; it halted part-way and the virtual disk went offline. The chassis was powered down that afternoon and every drive came in with its bay number written on the carrier. The two weak members were copied first, one of them across four passes, and the rest read straight through. The PERC's own description of the set was lifted off the images, the array assembled above them, and both volumes — the file share and the job database — brought up read-only. Everything was opened here before it went home.

Each of the weak members copied first2 volumes back on line

What helps, and what makes it worse.

First things to do

  • Power the chassis down as soon as a member drops out
  • Write the bay number on every carrier before removal
  • Send every disk, the failed one included
  • Say what the PERC has already been asked to do

What makes it worse

  • Starting or restarting a rebuild with a member missing
  • Importing or clearing a foreign configuration
  • Forcing a failed member back online in the menu
  • Re-creating the virtual disk to see if it returns

Answers before you commit to anything.

The rebuild stopped part-way. Is the array gone?

Usually not — but this is the point at which to stop touching it. Rebuilding puts every remaining member through a full sector read, the heaviest demand anyone can make of the oldest disks in the building, and it is generally the weakest that gives out under it. You are then two members down instead of one, with images still to be taken from both. Power the chassis down, leave the carriers in their slots, and let copies be made before anything else is tried.

The disks read as nothing on a PC. Are they damaged?

Almost certainly not. A SAS drive needs a SAS port, which an ordinary desktop does not have, and even on the right port a PC finds a RAID member carrying the PERC's own metadata rather than a filesystem it can mount. The same disks read perfectly well on the right equipment. If one has already been put on a PC and Windows has offered to initialise it, the only thing that matters is that nobody said yes.

Do you need the server, or only the disks?

The disks. Send every member as a bare drive with its bay number written on it — carriers, chassis and controller can all stay where they are, because the array's own description is read back from the images rather than from your hardware. Tell us which PERC it was and what it has already been asked to do: a rebuild, an import, a clear, a member forced back online. That history shapes the order of the work more than anything else you can give us.

How much is it, and how long does it take?

Servers and RAID sets start at £500 + VAT, and that band is published here rather than held back until you ring. The diagnosis costs nothing, the figure is fixed in writing before any chargeable work starts, and on most jobs there is nothing to pay unless the data is actually back. Four to six working days is normal for a multi-disk set, from the parcel landing to it going home on fresh media, and urgent work moves up the list.

If this sounds like yours, start here.

Disks out of a NAS → When a RAID 5 rebuild goes wrong → When ransomware has the files → Every device we handle →

Nothing gets worse while it sits switched off.

You pay nothing for the look. What comes back is a list — the files that opened, the ones that did not, and one fixed figure to finish the job. Leave it switched off until you have that in writing.

0800 6890668

Get it looked at Posting it to Oxford