Example scenario based on the kinds of failures we see. Not an individual client result.
Synology RAID 5: the rebuild that took the volume down
A 4-bay Synology NAS in RAID 5 starts beeping: one disk has failed and a second is reporting bad sectors. Someone swaps the failed disk and starts a rebuild. The heavy read load pushes the weak second disk over the edge mid-rebuild, and the whole volume crashes.
How we handle it: We take all four disks out, image each one sector by sector with hardware imagers (working around the damaged areas of the weak disk), rebuild the RAID 5 stripe order virtually from the images, then repair the Btrfs file system. The original disks are never written to.
Lesson: Never force a rebuild while a second disk is showing errors. Power down and call.

