My first in-prod corrupted hard drive problem

· databases · Source ↗

A Swiss biopharma ICT engineer traced a production MS SQL backup failure through EDR, VSS, and Windows corruption to a magnetically weakened HDD sector — recovered by a $90 tool that looked like a scam.

What Matters

  • HDD Regenerator recovered bad sectors by rewriting with strong magnetic patterns, forcing drive firmware to remap physically marginal sectors to spare pool.
  • SQL Server page checksums confirmed no silent data corruption post-recovery; only weakly magnetized sectors, not scratched platters.
  • The triggering event was a vendor technician’s SQL patch causing heavy I/O on audit pages untouched for years, exposing already-decayed sectors.
  • RAID would not have helped: silent page corruption replicates faithfully to every mirror before detection.
  • Dell shipped a replacement drive under warranty but explicitly declined any data recovery assistance despite active enterprise support contract.
  • EaseUS paid version failed; Windows dism /Online /Cleanup-Image /RestoreHealth detected corruption but could not repair it.
  • [HN: @jtchang] No mention of SMART diagnostics having flagged the drive beforehand — a key missing early-warning layer.
  • [HN: @benlivengood] Article’s RAID analysis is likely wrong; most RAID controllers begin resilvering or surfacing read errors during scrubs before silent corruption propagates fully.

Original | Discuss on HN