When Single-Parity Meets Its Hard Limit
You were warned RAID 5 could only lose one drive…
But nobody expects the second one to drop.
One minute the server is slow… the next minute the volume is offline, the controller is angry, and the RAID is red across the board.
It feels catastrophic.
But how bad it actually is depends entirely on what happened in the seconds leading up to that second drive failure.
This page explains what you’re seeing, why RAID 5 collapses the moment a second member drops, what NOT to touch (most important!), and what recovery still is possible — safely.
What You See
- Array shows Offline, Failed, or Incomplete after the second drive drops
- One disk marked Failed, another marked Offline, Foreign, or Missing
- BIOS may show “DG Missing”, “PD Offline”, or “Unexpected Foreign Configuration”
- OS sees no filesystem or block device
- Logs show read errors, timeouts, or “unrecoverable sector” events before collapse
- Reboot does not bring the virtual disk back
Why It Happens
1. Single-Parity = No Safety Margin
RAID 5 can recover from one missing drive — never two.
The moment a second disk drops, every stripe missing that sector becomes unrecoverable.
Ref: TN-R5-001 §1
2. First Drive Didn’t “Fail First” — It Failed Earlier
Often the first drive was degrading for weeks, and the “second failure” simply revealed the hidden problem.
Ref: TN-R5-001 §2
3. Latent Sector Errors Convert a 1-Drop Into a 2-Drop Event
Even if only one disk is officially failed, a single unreadable sector on a surviving disk produces the same effect as losing a drive entirely.
Ref: TN-R5-001 §3
4. Hot-Swap or Slot Event Mislabels Healthy Drives as Foreign
RAID metadata mismatches can make a perfectly good disk appear incompatible, causing the controller to eject or hide it.
Ref: TN-R5-001 §4
5. Power or Vibration Stress Causes Cascade Failures
Older arrays often collapse during rebuild or heavy IO when weak sectors finally surface.
Ref: TN-R5-001 §2, §3
What NOT To Do
- Do NOT force either disk online (rewrites parity blindly)
- Do NOT initialize or recreate the array
- Do NOT clear foreign configurations before cloning
- Do NOT swap disks to “see if it comes back”
- Do NOT try file-system repair utilities
- Do NOT apply firmware updates during failure mode
Every one of these actions overwrites original parity maps — the core evidence ADR needs to reconstruct data safely.
What You CAN Do
- Power down safely — prevent further sector damage
- Clone every member drive (failed, offline, foreign, healthy)
- Map each disk by slot → serial → WWN
- Export controller logs and config data
- Identify which drive failed first vs. which drive dropped second
- Reconstruct original stripe alignment offline
- Build a virtual RAID to test integrity before touching disk hardware
- Recover filesystem structures from binary-level evidence
When done correctly, many RAID 5 “two-disk failures” are only logically failed, not physically unrecoverable.
RAID Triage Center – RAID 5 Triage – RAID 5 Technical Notes