RAID 0 Has No Spare Half

The decode

RAID 0 splits every file across both drives, which is the whole point and the whole danger. Striping writes alternating pieces of each file to each member to double throughput — there is no duplication and no parity anywhere. That design buys speed by spending safety entirely: the array has less resilience than a single drive, because now two drives must both survive for any file to be whole.

One dead member therefore takes the entire volume, not its share. Because every file's pieces alternate between the drives, losing one member removes every second piece of every file. The surviving drive holds half of everything and a complete copy of nothing. That is why "the Windows drive failed" does not mean the Windows data is half-recoverable and the Mac data safe — both partitions live striped across both physical drives, and both are down until the failed member is imaged.

The recovery hinges entirely on the failed drive, which is the real work. The healthy member images trivially; the case stands or falls on getting a good image of the failed one. Its fault is diagnosed first — electronic, firmware, or mechanical — and imaged accordingly, because the reassembled volume can only be as complete as the worse of the two images. Everything else is reconstruction once that image exists.

Reassembly is virtual and exact, never done on the original drives. With both members imaged, the stripe is rebuilt from the images — stripe size, block order and member sequence read from the data and verified, since a wrong parameter yields plausible-looking corruption. The Windows and Mac partitions then emerge from the correctly reassembled volume, extracted to new media.

The completeness of the result tracks the failed member's condition. If the failed drive images fully, the whole striped volume reconstructs and both partitions come back intact. If parts of it never read, those gaps fall across files throughout both partitions — striping spreads the damage everywhere rather than confining it — and the affected files are reported honestly, named from the reconstruction.

The lesson is the one striping always eventually teaches. RAID 0 is storage with the safety deliberately removed, and it needs a backup more than a single drive does, not less. Whatever is recovered goes to redundant storage, and the striped array is not rebuilt without a backup behind it.

On the bench

Both members were imaged individually under hardware write-blockers — the healthy one directly, the failed one assessed first on PC-3000 UDMA and, being mechanical, opened in clean air under laminar flow for donor parts and imaged on DeepSpar Disk Imager with head mapping. The RAID 0 stripe was reassembled virtually from both images in UFS Explorer RAID, stripe size and order confirmed against the data, and the Windows and Mac partitions extracted from the reconstructed volume and hash-verified.

The outcome

Both members imaged, the failed one recovered mechanically, and the stripe reassembled virtually with both partitions extracted. The assessment is free and the quote is a single fixed figure inclusive of VAT; where a drive has to be opened, half of the parts and labour is payable up front and the balance falls due only on success — otherwise it is no recovery, no fee. The decode: a striped array has no spare half. One failed member takes the whole volume, so the recovery is really about imaging that one drive well — reassemble it correctly and both partitions come home together.

Our case files are written up from genuine enquiries our lab has handled for customers across Staines, Surrey and the surrounding area, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery approach our engineers apply to that fault, using the equipment listed.