A routine disk swap spiralled into a full array failure, days short of a filing deadline. Recovered without one write to the original disks.
A Newcastle accountancy practice lost its main server in the middle of what should have been a simple repair. The machine ran a four-disk RAID 5 — a layout that survives one disk failing, because the data is striped across the disks with parity that lets a missing disk be rebuilt. One disk had indeed failed, a replacement went in, and the rebuild began — but partway through, a second disk dropped out, and with two disks effectively gone the array fell offline. On it sat the firm's SQL databases and accounts data, only days from a filing deadline.
This is among the commonest and most dangerous RAID situations we meet. A failed rebuild is the exact moment most data is lost for good, since re-running it writes to the very disks you still need to recover from. We assessed each disk on its own. One was genuinely failing, carrying a band of unstable sectors; the others read fine but held an array left inconsistent by the aborted rebuild. From that point the cardinal rule is plain: attempt no further rebuild on the original disks.
Each member disk was pulled and imaged separately behind a hardware write-blocker. The failing one went onto a DeepSpar Disk Imager to recover the most good sectors possible without stressing it any further, while the PC3000 took the disks needing firmware-level attention. Reconstruction began only once a full set of images was in hand — entirely in software, never on the original hardware. Analysing the images gave us the parameters the controller had run: the stripe (block) size, the disk order and the way parity rotated across them. With that geometry right, we built a virtual array from the copies, the parity reconciled, and the file system mounted.
From the rebuilt array we verified the SQL databases and accounts files were complete and opened correctly, then returned them on fresh media.
It all came back intact and was returned comfortably ahead of the deadline — five working days start to finish, without a single write to the original disks. The lesson we press on every business running a RAID: when an array fails, stop, and let nobody attempt a rebuild until the disks have been imaged, because the rebuild is what most often turns a recoverable case 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.
Getting your data back begins with getting the device to us. Pack it up safely, pop your contact details inside, and send it over — once we’ve run the free diagnostic, we’ll confirm your exact price in writing before any work starts.
Posting it? A tracked, insured service is what we’d recommend. Rather drop it in? You’re welcome Monday to Friday, 9am to 5:30pm — just package the device up as above first.
Want a bit more detail first? Fill in the form with more about your issue and an engineer will review it and send you a custom quote.
We’ll be in touch shortly. For anything urgent, call 0191 406 1051.
Yes — any level, controller or failed rebuild. We image every member disk read-only, recover the parameters, then rebuild the array virtually from the copies.
From £500 plus VAT, with no fix, no fee on most jobs and a fixed quote up front; emergency round-the-clock service is available too.
No. A failed rebuild is the single most common cause of permanent loss. Stop, pull 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.