An IBM server's RAID 5, built on enterprise SAS drives, failed after a second disk dropped — taking the company's databases and file shares with it. We imaged every SAS member on the right hardware, allowed for their unusual sector format, and reconstructed the array offline to recover the data.
The array had lost a disk and kept running in a degraded state — unnoticed — until a second SAS drive failed and the RAID 5 dropped offline, taking the server's databases and shared files with it. Backups were out of date. Because a RAID 5 can only tolerate a single failed member, a second failure meant the array could no longer rebuild itself, and the right move was to stop, remove every disk, and bring them in rather than attempt an on-controller rebuild that could have made things worse.
SAS drives are enterprise hardware and can't be read on the ordinary SATA connections consumer drives use; imaging them takes the right SAS controllers and cabling. They also frequently store data in 520- or 528-byte sectors rather than the usual 512 — the extra bytes carry data-integrity information the controller uses to detect corruption. That formatting has to be recognised and handled correctly, or the reconstructed data comes out misaligned. On top of this sits the RAID 5 itself: data striped across the disks with distributed parity that lets one missing member be recalculated.
Each SAS disk was imaged individually on appropriate hardware. Members that read cleanly were cloned in full; any with weak areas were imaged adaptively, securing the healthy majority first. Working from images of every disk means the array can be reconstructed and verified without ever touching — or further stressing — the originals, and gives the fullest possible set of data to rebuild from.
The RAID 5 parameters — stripe size, parity rotation, disk order and start offset — were derived by analysing the raw images, using parity relationships and known file-system structures to confirm the map. The array was then rebuilt in software, with the drives' data-integrity sectors accounted for so the data landed correctly. Because images of all members were available, parity was used to check and correct rather than blindly trusted, so the data lost with the failed members was reconstructed wherever the surviving disks and parity allowed.
The recovered databases were validated and files opened across the shares to confirm they were intact, then everything was returned on fresh storage. About 98% came back, the small remainder corresponding to areas where two members were simultaneously unreadable. We stressed the point every degraded array makes: a RAID protects against one failure, not against neglect — a failed member needs acting on at once, and a real backup remains essential.
SAS imaging on appropriate controllers · handling of 520/528-byte data-integrity sectors · stripe, parity and disk-order reconstruction with PC-3000 RAID and Atola Insight · database validation. All imaging read-only, work carried out in-house in Belfast.
Send it to us for a free, no-obligation diagnostic. We’ll tell you what can be recovered and put a fixed price in writing before any work starts — and on most jobs, if we can’t get your data back, there’s nothing to pay. Post your device in, or drop it to us by appointment.
Yes — any make and RAID level. We image SAS members on the right hardware, handle enterprise sector formats, and reconstruct the array in software from the copies. Send every disk, labelled with its slot order, including any that failed.
Not necessarily. A second failure stops the array rebuilding itself, but if the failed disks can be imaged — even partially — the array can often still be reconstructed offline with most of the data intact. Don't let the controller attempt a rebuild.
RAID and server recoveries start from £500 plus VAT, with a fixed written quote after a free diagnostic. Physical repairs carry a deposit toward parts and bench time.