TN-R6-003 — Latent Sector Errors Under Dual-Parity Stress

0. Purpose and Scope

This note explains why RAID 6 rebuilds expose latent sector errors more aggressively than RAID 5.

1. Rebuild Read Amplification

• Reed–Solomon parity requires additional survivor reads.
• Each stripe exercises all remaining members.
• Cold data regions are fully scanned.

2. Latent Errors as Effective Drive Loss

• One UNC consumes one parity symbol.
• Second UNC at same stripe exceeds solvability.
• Controller halts to avoid parity corruption.

3. Practical Outcome

RAID 6 rebuilds stall not because parity is weak, but because parity math is being protected.