Data Recovery Case File · NAS & RAID · The Safe Migration

"I don't want to risk formatting the good drive": a QNAP rebuild, paused at exactly the right moment

Not every case in this archive starts with a disaster; this one starts with a near-miss that never happened. A six-year-old QNAP TS-453A running two mirrored 4TB WD Reds lost one drive to age. The owner did the responsible things — bought two replacements, planned to migrate — and then stopped at the edge of the instructions, recognising that one misread prompt during a rebuild could format the only good copy of everything. Her enquiry asked, in effect: can someone who does this daily do it for me?

The enquiry

“Recently one of the drives became faulty, so I've purchased two replacement drives and attempted to rebuild the RAID and copy the data from the good drive across. I'm unsure how to do this, though, and don't want to risk formatting the other drive and losing all of my data — so I need some help. Could you do this remotely, or would I have to send the QNAP and the drives to you? And what would it cost?”

Why her fear was well-founded — and why we said no to remote

Her instinct deserves validating in print: the rebuild window is the single most dangerous period in a mirror's life. The array is running without a safety net; the NAS interface is offering initialise-and-migrate prompts whose wrong answer is a format; and the surviving drive — six years old itself, sibling of one that just died — is about to be asked for its hardest week of work, a full-copy marathon of exactly the kind that finishes ageing drives elsewhere in this archive. Her second question got a straight answer too: no, not remotely — not because it's impossible to click buttons over a screen-share, but because remote hands can't do the one thing that makes the operation safe, which is taking a write-blocked image of the good drive before anything else is allowed to happen. So: send everything in.

The migration, done in the safe order

Order was the entire service. First, the surviving member was imaged in full on the TaskForce 2 — write-blocked, hash-verified — so that from minute one, her data existed twice and every later step became reversible. (The image also gave the drive its physical: it passed, but with the early signs of age that made this whole exercise timely.) Only then were her two new drives built into a fresh, healthy mirror in the QNAP, the data loaded onto it from the verified image, and the shares checked over against the original's own accounting — folder counts, sizes, spot-opened files. The faulty drive was returned labelled as scrap; the old good drive went back untouched and unformatted, with our suggestion that it live on a shelf as a frozen archive copy of this moment — a free backup her caution had already paid for.

Outcome

A degraded array retired without a byte lost, a new mirror running on new drives, and a fixed price she knew before shipping anything. The page's point is the category it represents: recovery labs are not only for the aftermath. The cheapest jobs we do are the ones commissioned before the disaster — assisted migrations, pre-emptive imaging of ageing drives, safe hands for the operations whose instructions end in "are you sure?". If a storage task has you hovering over a prompt you don't fully trust, that hesitation is expertise in its earliest form. Listen to it, as she did.

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.