RAID Parameter Recovery: Stripe Size, Layout, Order and Offset
A RAID set without its controller is just N disks. Recovery means proving the five parameters from the data itself, not guessing them.
The five parameters
Member order, stripe size, parity layout (left/right, sync/async), parity rotation and the data offset at the start of each member. Any wrong assumption produces a filesystem that looks almost valid — which is why 'it mounts but files are corrupt' is a common failure mode.
Evidence beats guessing
Parameters are scored against self-consistency: filesystem superblocks, MFT records, and the distribution of parity blocks across members. Two candidate sets can both 'look' plausible; the scoring has to separate them.
Never rebuild first
A controller rebuild writes to the members. If one member is weak, the rebuild can finish the job of destroying the set. Image every member read-only, then assemble virtually.
FAQ
Can RAID 5 be recovered with two failed disks?
Only at file level and only if the failed members can be imaged partially; redundancy is exhausted.
Do I need the original controller?
No, but its metadata (or a same-model config dump) shortens the work considerably.
How long does it take?
Typically 1–5 business days depending on capacity and member count.
Boundary: more missing members than the redundancy allows, or overwritten parity, limits recovery to file-signature salvage.
Free evaluation in our Shenzhen lab; written quotation after inspection; no recovery, no fee. Mail-in service available for Hong Kong and overseas clients.