Specialist · recovery from servers
Failed platters are seldom the cause. Far more often it is state: a card that no longer believes its own disks, an array one prompt away from oblivion, writes still stranded in cache. We rebuild from images — the original card is not needed — and return the volumes, virtual machines and databases your business runs on.
Speak to an engineer about it
0800 6890668
| The problem | What that tells you | First step |
|---|---|---|
| Dell PERC: Foreign Configuration Detected | What the disks record disagrees with what the card remembers — it has halted as a precaution, not because data is lost | Pick neither option |
| 1786 — Drive Array Recovery Needed (HP) | Either a rebuild is outstanding, or one stopped part-way | Switch off; take images before anything else |
| Orange bay lights following a power cut | Half-finished writes and cache trouble; the disks still hold the layout | Resist forcing any disk back online |
| It starts up, but the volume has vanished | Card state or file-system damage above untouched data | Good chances of recovery |
| The RAID card itself has failed | The layout, RAID 5 parity rotation included, lives on the disks | An identical card is unnecessary |
| Nothing mounts after the ‘recovery’ — guests or databases | Getting files back is not enough; they must be consistent | Our VMware and database pages cover this |
As soon as it lands with us, your device gets its own case number. An engineer works out the fault, says what can genuinely be pulled off, then puts one fixed price in writing — no charge for diagnosis, no obligation, and no chargeable work until you approve it.
Switch off, mark which bay held which disk, answer nothing. Every member is imaged separately on dedicated hardware — the ones flagged failed included — before anything is interpreted.
RAID level, stripe width, member order and — on RAID 5 and 6 — parity rotation come out of the DDF or RIS records in the images, and the array is assembled in software. Losing the card costs you nothing.
Whatever sits above the array — a VMFS datastore, NTFS, ReFS, ext4 or XFS — is put right on the rebuilt copy, and then what matters gets tested: virtual machines boot, databases attach, shares open.
You sign off a complete list of the recovered files first; only then does the recovery fee fall due. Everything returns on fresh media, postage paid, and the job stays open until you confirm the files open at your end.
Why that two-option prompt has the reputation it has: Import lays the foreign metadata’s view of the array over the one the card is running; Clear wipes the DDF map off the disks altogether. Each is one keystroke, neither can be undone, and both stop being dangerous once every disk exists as an image. Imaging first is not caution here — it is the whole plan.
The mains failed, the UPS ran dry, and at restart the Smart Array wanted choices nobody could make safely. All four drives imaged, the RIS metadata parsed, delayed parity respected — and the practice-management database mounted off the rebuild, whole to the last write the cache flushed before the power went.
Usually not. Most server failures come down to array metadata, controller state or file-system damage sitting on healthy disks. Shut it down, stop clicking through prompts, and the odds stay firmly on your side.
The Dell PERC is seeing array metadata on the disks that disagrees with what its own memory holds — after a board change, drives being moved, or a controller fault. The halt is a precaution. Either option it offers you can wipe data; image first and the choice stops being dangerous.
Yes. Each member carries the layout with it: PERC and MegaRAID lay down standard DDF metadata, HP its RIS blocks. Reconstruction happens on the bench from images, and no matching controller is needed.
Both — and the difference counts. Pulling back a VMDK or MDF file only gets you halfway; the rest is persuading the hypervisor or database engine to mount it. That layer is what our VMware and database pages deal with.
Powering a damaged device up again costs you data. Start a case first; diagnosis is free whatever you decide.