A four-disk RAID 5 that the operating system stopped seeing entirely. The disks were fine. The array was fine. The controller was the thing that had failed — and that is a distinction worth understanding before you let it try to fix itself.
← All case files · from £500 + VAT
A business RAID 5 array, four disks, holding financial records and project files. No backup. The operating system simply stopped recognising the volume — not degraded, not rebuilding, just absent.
The cause was firmware corruption on the RAID controller itself, which had left the card unable to communicate coherently with its own members.
The card could no longer read its own configuration, so it presented no array at all.
of their own — not enough to have failed outright, but enough to make a controller-driven rebuild lethal.
, so no automatic rebuild was ever going to succeed.
This is the single most useful thing on this page. People believe the RAID card holds their data. It does not. The disks hold the data; the card holds the map. When the card corrupts, the data is still sitting on the platters exactly where it was — and the one action that will actually destroy it is letting a confused controller write a new map over the top.
A RAID 5 with two flaky members is one bad rebuild away from being a forensic job. That is not a hypothetical; it is the most common way we see arrays destroyed.
The array was reconstructed from the clones and the data returned. The controller was never repaired, because it never needed to be — it was not holding anything.
What determines the outcome on a job like this is not the controller at all. It is how completely the weakest member images. RAID 5 tolerates one missing disk by calculating it back from parity. Lose a second member's sectors in the same stripe and that stripe is gone for good, which is why the two unstable drives were the entire job and the corrupted card was a footnote.
If your RAID controller starts behaving oddly, stop. Do not reseat drives to "see if it helps". Do not accept the prompt to initialise, import a foreign configuration, or rebuild. Do not update the controller firmware in the hope it clears.
Photograph the drive order first — which disk was in which bay — then power it down and label them. Disk order is one of the variables that has to be solved in the lab, and if you already know it, you have saved yourself real time and real money.
Drop the drive at our Quayside reception, or post it to us — it costs nothing to find out what happened. You get a written figure from the fixed bands before any work begins.