Data Recovery Case File · Professional Photography · The Mirror's Promise

Three seconds of power, two complete copies, one correct customer theory: a dual-drive mirror's enclosure failure decoded — and time-sensitive wedding work delivered back

The enquiry arrived urgent, professional, and — rarest of all — correctly reasoned end to end. A working wedding photographer's SanDisk Professional G-RAID Mirror, 36TB, bought less than a year earlier and configured as RAID 1: "It suddenly stopped mounting on my iMac and now powers on briefly (for about 3–5 seconds) before shutting off. It does not appear in Disk Utility or Finder." Ports, cables and machines all eliminated. The cargo: "critical and time-sensitive wedding photography files." And then the reasoning this page exists to ratify: "I understand RAID 1 should mean each internal drive contains a full, identical copy of the data. I am therefore hoping that at least one of the [drives survives]." His understanding is exactly right — and the three-second power signature says the odds are even better than he hoped, because the thing dying in that box was very probably neither of his copies.

The decode: what three seconds of power convicts, what the mirror promises — and the warranty question answered straight

The power signature first: a unit that starts, runs three to five seconds, and shuts itself off is performing a protective abort — its power stage or controller board detecting a fault during start-up and cutting out rather than running wrong. That behaviour lives in the enclosure's electronics: the supply, the RAID controller, the bridge — the box around the drives, not the drives themselves, which in these cases typically never even complete spin-up before the plug is pulled on them from inside. On a nearly-new unit with no drop, no surge and a sudden onset, the failing-box-healthy-drives split is the strong default — and it means his "at least one copy" hope was, if anything, conservative: the likelier truth was both. The mirror's promise, ratified formally: RAID 1 writes everything twice — each member drive carries a complete, independently readable copy of the archive, and that is precisely the configuration's whole purpose: any single point of failure (a drive, or as here, the enclosure that housed them both) leaves a full copy standing. His enquiry stated the principle textbook-correctly, and the assessment's job was simply to confirm the real world had honoured it. One professional caution the mirror's promise carries in small print, added because his conduct already respected it: the two copies are only independent while nothing reconfigures them — enclosure swaps, new-unit migrations and helpful setup wizards can initialise members they don't recognise, and the safe order after any RAID enclosure failure is always drives out, read directly, decisions later. The warranty question, answered because every under-a-year case deserves it straight: a unit this new is almost certainly still covered — and the warranty will replace the box. It will not, and never does, recover the work inside it; worse, warranty returns typically require sending the unit (drives and all) back, and units come back replaced, blank, or both. The professional order is therefore fixed: data first, warranty second — recover the archive from the members, then let the empty enclosure claim its replacement with nothing left to lose.

The recovery — the promise, kept at pace

The members came out of the failing box and were assessed independently — both drives healthy, spinning cleanly on direct native connection, the enclosure's protective abort confirmed as the sole author of the outage — and the mirror's promise was kept in the most literal way available: the primary copy read directly and in full, the second member serving as live verification twin, every byte's twin checked against it. The wedding archive mounted whole: the time-sensitive deliveries located, opened and verified first — galleries confirmed image by image at the working-professional standard — with the full 18TB estate following, delivered on fresh independent media in duplicate at the turnaround pending client work demands. The enclosure went back to him separated from its former cargo, cleared for its warranty journey; the report ratified his reasoning in print (RAID 1 functioned as designed; both copies intact; enclosure electronics sole fault); and the delivery note carried the one structural addendum his setup needed: a mirror in one box survives a drive or a box — it does not survive a desk fire, a theft or a bag; the archive's third copy belongs in a second postcode, and for a working photographer, so does the client-delivery set.

Outcome

The weddings delivered, the mirror's promise honoured, the customer's theory ratified word for word — and the filings for every professional running paired storage. Read the three-second shutdown as the box, not the drives: protective aborts live in enclosure electronics, and nearly-new units failing this way are overwhelmingly housing casualties over healthy members — stop power-cycling it (each abort is the fault re-firing) and get the drives read directly. Trust RAID 1's promise, and protect it: each member is a complete copy until something reconfigures it — no new enclosures, migrations or setup wizards until the data's out. Run warranty in the right order: the cover replaces hardware, never work — recover first, claim second, and never ship loaded drives to a returns process. And note what made this case fast: a correctly-reasoned enquiry — configuration stated, elimination done, stakes named. The mirror did its job. The box did its dying. The photographer, start to finish, did his thinking — and the weddings never knew a thing.

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.