Data Recovery Case File · NAS & RAID · Vendor Support's Homework

"Access it via Linux, or repair the installation": a Synology support ticket's two options, weighed and one executed

This customer arrived unusually well-briefed, because he'd done the vendor process properly: extensive discussion with Synology support about his dead two-bay unit, ending in a genuinely useful diagnosis — no hardware fault found in the enclosure or either mirrored WD Red, the problem likely in the on-disk partitions or the DSM system installation — and two suggested routes: recover the data via a Linux installation, or fresh-install onto a new disk and use that installation to "correct" the affected ones. He had the diagnosis and neither the setup nor the appetite to gamble on execution. Wise on both counts.

Grading the two options

Option two first, because it's the one this page exists to caution against. "Fresh install, then use it to correct the affected disks" is sound vendor logic for restoring an appliance — and a gamble when the disks being corrected hold the only copy of the data, because every verb in it writes: installation writes system partitions, "correcting" writes wherever the correction logic decides, and a repair process meeting corruption it doesn't fully understand can standardise its way through the very structures a recovery needs. Vendor support's remit is a working unit; ours is the data; the options rank differently depending on which you're protecting. Option one — read the disks' Linux-native contents directly — was the right instinct wearing DIY clothes, and its professional form adds the two things the DIY version lacks: a safety net, and the full anatomy. Both members were imaged write-blocked on the TaskForce 2 before anything else existed in the plan — from that moment, no subsequent step could cost him anything — and the assembly then honoured the appliance's real layering, familiar from every Synology in this archive: system partitions (where the corruption lived, exactly as support had deduced), the mirror container assembled from the images with its two members cross-checked sector against sector, and the data volume inside, structurally sound behind its broken front-of-house.

Outcome — support's diagnosis, vindicated

The extraction confirmed the ticket's conclusion in full: hardware innocent, system installation corrupted, data untouched — his shares came off the assembled volume complete, folder trees and file dates intact, verified and delivered on a plain drive. The old members went back unmodified, one flagged as a ready-made cold archive of this exact moment. Credit where due closes the page: vendor support that says "we found no hardware fault, the issue is likely here, and here are your routes" has done honest, competent triage — the single upgrade this case argues for is in the routing, not the diagnosis. When a suggested remedy involves installing or repairing onto disks that hold the only copy, that's the handover point — the moment the job stops being appliance restoration and becomes data recovery, whoever's letterhead the advice arrived on.

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.