A dead server runs two emergencies at once — the data and the downtime — and they pull in opposite directions. Downtime says try everything now; the data says touch nothing. Twenty-five years of Glasgow server jobs have taught us exactly where that line sits: protect the disks completely, then move as fast as the imaging allows. Firms survive days of downtime. They don’t survive the payroll database being rebuilt over.
Power the box down cleanly if it’s still half-alive; if it’s already dead, leave it dead. Photograph the drive bays before anything moves, pull nothing hot, and above all start no rebuild — not from the controller BIOS, not from the vendor utility, not “just to see”. Then ring us. The first ten minutes of that call are free triage: what failed, what’s safe, what the realistic timeline is, and whether the emergency track is worth its premium for your situation.
What you should never do is the thing that feels most productive: reinstalling the OS on the array “to get something running”. That single act does more damage than the original failure.
The pattern rarely varies. A disk fails quietly in a Dell PowerEdge or HP ProLiant and nobody notices for months, because RAID does its job silently. Then a second disk falters, or a well-meant rebuild loads the survivors past their limits, or the PERC or Smart Array controller itself dies and offers to “initialise” on its way out. SAS or SATA, hardware or software RAID, Windows Server, VMware host or Linux box — the bench treatment is identical: every member imaged individually, the array solved virtually, the volume and its databases repaired against copies.
Hot spares get imaged too. More than once the “spare” turned out to hold the only clean copy of a stripe.
A server recovery isn’t finished when files exist — it’s finished when the business runs. So the output is shaped for restoration: Exchange stores and SQL databases verified mountable, VM files delivered ready for a hypervisor, share structures and permissions preserved where the file system allows, and the recovered set handed over on encrypted media your IT can restore from directly. If your engineers want the play-by-play for their incident report, they’ll get it in plain English.
Multi-disk work starts at £500 + VAT and is quoted to the job, because every member disk is imaged individually before the array itself is touched — more disks, more careful hours. Every job starts with a free diagnostic and ends the same way it was quoted: the figure goes in writing before a single sector is read, and on most jobs there is nothing to pay unless your data comes back. No hourly meter, no surprise “evaluation fee”, no percentage of what the files are worth.
Rebuild gone wrong, or a volume that vanished after a controller swap? The guide to how each RAID level fails covers exactly these patterns.
Fast, safely: imaging begins the day the disks arrive, and the emergency track runs evenings and weekends when the situation justifies it. What we won’t do is skip the imaging to save hours — that trade sacrifices the data to the clock. Call first and we’ll give you a realistic timeline before you commit to anything.
Usually not — a failed rebuild overwrites some regions and leaves others, and the pre-rebuild state can often be reconstructed from what each member still holds. Honesty matters here: results after a rebuild attempt are frequently partial. Stop all further attempts, note what was done and when, and send everything including any replaced disks.
Yes — recovering the volume is half the job on a business box. Exchange EDB stores, SQL Server MDFs and the like are checked, and where the failure left them dirty or corrupt they’re repaired at page level from the recovered copies, so what you get back actually mounts rather than merely exists.
Drives are enough, and lighter to ship — pull them, number the bays, pad them individually. Send the whole chassis only if removing disks is impractical or the controller configuration is a mystery worth preserving. Either way, include the failed and replaced disks too; “dead” members often still hold the deciding stripe.
Two honest routes, no vans. Hand the device in at Tay House, 300 Bath Street — right at Charing Cross, two minutes off the M8 — Monday to Friday, 9am–5:30pm. Or wrap it well and send it by insured, tracked post from Paisley, Ayrshire, the Highlands or anywhere else in the UK. Either way the work is done in-house by our own engineers, with a documented chain of custody, and your data never leaves the UK.
Photograph the bays, pull the disks, start nothing — and call before anyone else touches it. Server data recovery for Glasgow businesses: imaged like evidence, returned like a Monday-morning plan, brought in by hand or insured post.