An Edinburgh financial services firm ran its primary storage as a four-drive RAID 0 on an IBM server — maximum speed, zero redundancy — with no external backup behind it. When one drive’s read/write heads failed mechanically, the whole striped array went dark at once: client records, financial data and operational files, all inaccessible because a quarter of every file lived on the dead disk.
RAID 0 leaves no parity to lean on, so the failed member itself had to read again. Its damaged head assembly was replaced with donor parts from a matching drive model under clean-air conditions, the platters checked for contamination, and the repaired drive imaged sector by sector with the weak regions handled last. The three healthy members were cloned untouched before any of that began — house rule. From the four images, the stripe geometry (order, block size) was reconstructed and the array rebuilt virtually; the NTFS file system on top was then repaired to restore its directory structure, and extraction ran priority-first: the financial records and client data the firm named at intake, then everything else.
97% of the array’s data went home in five days, including every essential financial and operational file, delivered on a new external RAID with verification against the client’s priorities. The missing few percent lay in the dead drive’s worst sectors — counted and disclosed, not glossed. The firm left with two recommendations it took: a real backup regime, and a migration off RAID 0 to a fault-tolerant level for anything that matters.
One dead disk in a RAID 0 takes everything with it — and recovery means bringing the dead disk back to readable, then rebuilding the stripe mathematics from images. The RAID recovery page covers the discipline. The first step never changes: a free diagnostic and a fixed written quote before anything is at stake — or call 0131 202 0491 and describe what happened.