Data Recovery Case File · NAS & Network Storage · The Layout Is on the Disks
You Do Not Need the Card, You Need the Parameters
His enquiry diagnoses the obstacle correctly and draws the wrong conclusion from it. Professional audio files on drives from an old machine: "I have found out the drives were set up as striped. That's all I know. Have tried plugging into my new computer but nothing is recoverable, probably as I don't have the array controller." He is right that the controller is missing and wrong that it is required — because the arrangement it managed is recorded on the disks it managed.
| Media | Multiple hard drives from a striped array — original hardware controller unavailable; members presenting individually without a usable volume |
| Reported situation | Drives from an older machine used for professional audio work · configured as a striped set without redundancy · original array controller no longer available · drives connected individually to a replacement machine · no content accessible · configuration details not known to the owner |
| Fault class | Striped set without its original controller — layout recoverable from member metadata or determinable analytically; controller hardware not required |
| Equipment used | Members imaged individually write-blocked before any assembly · array metadata read from the disks themselves · stripe size, member order and offset determined analytically where metadata was absent · volume assembled offline from images · project path structure preserved in delivery |
The decode: where the configuration actually lives
What a hardware controller does: presents several physical disks to the computer as one volume, by splitting every write across them according to a defined pattern — how large each piece is, which disk comes first, and where on each disk the data begins.
Where it keeps that pattern: in two places. In its own memory, and as metadata written onto the member disks themselves. The second copy exists precisely so that an array survives its controller being replaced — a card that dies can be swapped for another and the new one reads the arrangement off the disks.
So what follows for him: the parameters are on the drives. The card is a convenience for running the array, not a requirement for reading it, and a reconstruction reads that metadata directly rather than needing hardware that no longer exists.
What happens where the metadata is missing or unreadable: the parameters are determined analytically. There are a limited number of plausible combinations — stripe sizes come in a small set of standard values, member order is a permutation of the disks present, and offsets are conventional. Candidate combinations are tested against the content until one produces coherent filesystem structures and files that open.
Why that verification step matters: a wrong stripe size or a wrong member order produces a volume that mounts and contains nonsense. The right answer is the one where files reconstruct and open, not the one that merely assembles.
What is genuinely required, and this is the part at risk: every member, unaltered. Connecting them to a different array controller is the danger — a controller meeting disks whose metadata it does not recognise frequently offers to initialise them, which writes fresh metadata over the record of the original arrangement.
Why the folder structure matters more than usual here: professional audio projects reference their source material by path. A recovery that returns audio files without their arrangement returns a library rather than a project, and the structure is a primary deliverable rather than a convenience.
What he should establish: how many drives there were, and that he has all of them. A stripe missing a member is missing a fraction of every file.
On the bench
Members were imaged individually write-blocked before any assembly, and array metadata read from the disks themselves — a hardware controller writing the layout onto its members as well as holding it internally, precisely so an array survives the controller being replaced. Stripe size, member order and offset were determined analytically where metadata was absent, candidate combinations tested against content until filesystem structures resolved coherently. Project path structure was preserved in delivery.
The outcome
Every member imaged before assembly, the layout read from the disks and confirmed against content, and the project structure preserved. Free assessment, one fixed written figure including VAT; where a drive has to be opened, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The decode: you do not need the card. A controller writes the array layout onto the member disks as well as keeping it internally, exactly so the array survives the card being replaced — so the parameters are on your drives.
Array disks without the controller that ran them
Don't connect them to a different array controller — that's the real risk, because a card meeting disks whose metadata it doesn't recognise frequently offers to initialise them, writing fresh metadata over the record of your original arrangement. The card itself isn't needed: controllers write the layout onto the member disks as well as holding it internally, precisely so an array survives its controller being replaced, and a reconstruction reads that directly. Where the metadata is gone, the parameters can be determined by testing candidate combinations until files actually open. What you do need is every member, unaltered — so establish how many there were.
Don't attach them to another card — call Oxford Data Recovery on 01865 593000; members imaged before assembly, layout read from the disks, parameters confirmed by whether files reconstruct and open.
Request a quote online →
Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.