8 × 2 TB in RAID 5: the numbers
| Usable space | 14 TB (12.7 TB in Windows) |
|---|---|
| Raw space | 16 TB |
| Used for protection | 2 TB |
| Efficiency | 88% |
| Drives that can fail | Any 1 |
| Rebuild at 150 MB/s | 3 h 42 min |
| Rebuild at 50 MB/s (busy) | 11 h |
| Data read during rebuild | 14 TB |
| Read error risk, desktop drives | 67% |
| Read error risk, NAS drives | 11% |
Rebuild
Replacing a failed drive means writing a full 2 TB drive. At 150 MB/s with no other load that takes about 3 h 42 min; on a busy array running at 50 MB/s, about 11 h. The rebuild reads every surviving drive: 14 TB in total.
Read errors
Reading 14 TB, the chance of at least one unreadable sector is 67% with desktop drives (1 per 1014 bits) and 11% with NAS drives (1 per 1015). Nothing is left to correct it, so the affected file, or on some controllers the whole rebuild, fails. These are worst-case datasheet rates; real drives usually do better.
Our take
With 8 drives sharing one parity, a second failure during the rebuild gets more likely. RAID 6 gives 12 TB with two-drive protection. RAID 10 with the same drives gives 8 TB, faster writes and shorter rebuilds.