The Mechanism
When a drive fails in a RAID 10 array, the array becomes degraded, but performance remains relatively high compared to RAID 5 or 6. To fix it, you insert a new drive to trigger a rebuild.
During a RAID 10 rebuild, the controller does not need to read every single drive in the system to calculate mathematical parity. Instead, it performs a direct, sequential read from the one surviving twin drive in the mirrored pair to clone its data onto the new replacement drive.
The Impact on RAID 10
Because the rebuild stress is isolated entirely to that single twin drive, that specific drive is pushed to its mechanical and electrical limits. If all the drives in the array were purchased at the same time and have the same wear level, the surviving twin is highly vulnerable. If it suffers a head crash or a sector failure mid-rebuild, the mirrored pair dies, and the entire RAID 10 array fails.
