Data Recovery Case File · Solid State & Flash · The Misleading Name
Striping Doubles the Number of Ways to Lose It All
His enquiry describes a configuration whose name misleads almost everybody who uses it. Two 1TB drives "set up in RAID 0, and one of them seems to have failed. The configuration was done in the firmware prior to installing the operating system." The name places it alongside arrangements that protect data. It is the only one that protects nothing — and one drive failing does not cost half the contents, it costs all of them.
| Media | Two 1TB NVMe solid-state drives in a striped configuration without parity — one member failed; volume unavailable |
| Reported situation | Two solid-state drives configured as a striped set in system firmware · configuration performed prior to operating system installation · one member apparently failed · volume no longer available · contents required |
| Fault class | Striped set with a failed member — no redundancy present; reconstruction requiring both members and the original stripe parameters |
| Equipment used | No firmware reconfiguration or rebuild attempted · both members imaged individually write-blocked · stripe size and member order determined from controller metadata and content analysis · volume assembled offline from images · output restored to separate hardware |
The decode: what striping does, and why the name causes the misunderstanding
What the configuration actually does: splits every file across both drives in alternating pieces. A file is written a chunk at a time — some to the first drive, some to the second, back and forth. Neither drive holds a complete anything. The point is speed: two devices read and write simultaneously, so throughput roughly doubles.
Why one failure loses everything: because every file has pieces on both drives. Lose one and every file is missing alternating chunks throughout its length — not the second half, not some files, every file, in pieces, uniformly. The surviving drive holds half of everything, which is a great deal of data and nothing usable.
Why the name misleads: it shares an acronym with configurations built to survive a failure, and the number suggests a level within a scale. People reasonably infer that having any of them is safer than having none. The opposite is true here: with two drives striped, the probability of losing the volume is roughly double that of a single drive, because either failing takes it.
Why solid-state members make it worse in one specific way: these devices frequently fail abruptly and completely rather than degrading, so there is often no warning at all — and the configuration provides nothing to catch it.
What recovery actually involves: both drives, including the failed one. The failed member has to be made readable first, then both are imaged, and the stripe is reassembled — which requires knowing the stripe size and the member order. Get either wrong and the reassembled volume is meaningless, so those are determined from the controller's metadata and confirmed by analysing the content rather than guessed.
Why the firmware configuration matters: the parameters live in metadata written by the controller, so they are recoverable — but it also means the set must not be reconfigured or reinitialised there. Firmware utilities offer to rebuild or recreate a set, and recreating one writes fresh metadata over the record of how the original was arranged.
What to do afterwards: if the speed genuinely mattered, keep the arrangement and back it up properly — a striped set is a working volume, not a home for anything singular. If it did not, run the drives separately.
On the bench
No firmware reconfiguration or rebuild was attempted — utilities offering to recreate a set write fresh metadata over the record of the original arrangement. Both members were imaged individually write-blocked, including the failed drive, since striping distributes every file across both in alternating pieces and neither holds a complete file. Stripe size and member order were determined from controller metadata and confirmed by content analysis, either being wrong rendering the assembled volume meaningless. The volume was assembled offline and restored to separate hardware.
The outcome
Both members imaged individually, the stripe parameters determined from metadata and confirmed against content, and the volume assembled offline. Free assessment, one fixed written figure including VAT; where a chip has to be removed, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The decode: striping splits every file across both drives in alternating pieces, so one failing leaves every file missing chunks throughout — all of it, not half. The name places it alongside configurations that protect data, and it is the one that protects nothing.
Striped set with a failed member
Don't let the firmware rebuild or recreate the set — that writes fresh metadata over the record of how the original was arranged, which is what a reconstruction needs. Keep both drives, including the failed one, because striping splits every file across both in alternating pieces and neither holds a complete anything: the survivor has half of everything, which is a lot of data and nothing usable. Worth knowing the name misleads. It shares an acronym with configurations built to survive a failure, so people reasonably assume it offers some protection, when in fact having two drives striped roughly doubles the chance of losing the volume because either one failing takes it.
Don't rebuild in firmware — call Guildford Data Recovery on 01483 901310; both members imaged individually, stripe size and order determined from metadata and confirmed against content, assembled offline.
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.