A routine disk swap turned into a full array failure, days before a filing deadline. Recovered without a single write to the original disks.
A Bristol accountancy practice lost its main server during what should have been a routine repair. The box ran a four-disk RAID 5, an arrangement that can survive a single disk failing because the data is spread across the disks with parity that allows a missing disk to be rebuilt. One disk had indeed failed, a replacement was fitted, and the rebuild was started — but partway through, a second disk dropped out, and with two disks effectively gone the array went offline. On it were the firm's SQL databases and accounts data, days from a filing deadline.
This is one of the most common and most dangerous RAID scenarios we see. A failed rebuild is precisely the moment most data is lost for good, because re-running it writes to the very disks you still need to recover from. We assessed each disk individually. One was genuinely failing with a band of unstable sectors; the others were readable but carried an array left inconsistent by the aborted rebuild. The cardinal rule from here is simple: do not attempt another rebuild on the original disks.
Each member disk was removed and imaged individually through a hardware write-blocker. The failing disk was imaged on a DeepSpar Disk Imager to recover the maximum number of good sectors without stressing it further, and the PC3000 handled the disks that needed firmware-level attention. Only once we had a complete set of images did the reconstruction begin — entirely in software, never on the original hardware. By analysing the images we recovered the parameters the controller had been using: the stripe (block) size, the order of the disks and the way parity was rotated across them. With that geometry correct, we assembled a virtual array from the copies, the parity reconciled, and the file system mounted.
From the reconstructed array we confirmed the SQL databases and the accounts files were complete and opened correctly, then returned them on fresh media.
Everything came back intact and was returned comfortably before the practice's deadline — five working days end to end, and not a single write made to the original disks. The lesson we stress to every business running a RAID: when an array fails, stop, and do not let anyone attempt a rebuild before the disks have been imaged, because the rebuild is what most often turns a recoverable situation into a lost one.
DeepSpar DDI · PC3000 — imaging and recovery carried out in-house. Every job is imaged before any recovery work begins, and the original media is never written to.
Send us your device for a free diagnostic, and tell us a little about what happened — an engineer will review it and confirm your exact quote in writing before any work begins.
Recovering your data starts with getting the device to us. Pack it safely, add your contact details, and send it over — after we run a free diagnostic, we’ll confirm your exact price in writing before any work begins.
Posting it in? We recommend a tracked, insured service. Prefer to drop it off? You’re welcome Monday–Friday, 9am–5:30pm — please still package the device as above.
If you need more information on our data recovery service, fill out the form with more detail about your issue and an engineer will review it and give you a custom quote.
We’ll be in touch shortly. For anything urgent, call 0117 332 1137.
Yes — any level, controller or failed rebuild. We image every member disk read-only, recover the parameters, and reconstruct the array virtually from the copies.
From £500 plus VAT, no fix, no fee on most jobs, with a fixed quote up front; emergency, round-the-clock service is available.
No. A failed rebuild is the most common cause of permanent loss. Stop, remove the drives labelled with their bay order, and send them to us.
Start with an instant online quote, or call and talk it through with us first. You'll have a clear, fixed price before any work begins.