3 × 14 TB in RAID 5: the numbers
| Usable space | 28 TB (25.5 TB in Windows) |
|---|---|
| Raw space | 42 TB |
| Used for protection | 14 TB |
| Efficiency | 67% |
| Drives that can fail | Any 1 |
| Rebuild at 150 MB/s | 25 h |
| Rebuild at 50 MB/s (busy) | 3.2 days |
| Data read during rebuild | 28 TB |
| Read error risk, desktop drives | 89% |
| Read error risk, NAS drives | 20% |
Rebuild
Replacing a failed drive means writing a full 14 TB drive. At 150 MB/s with no other load that takes about 25 h; on a busy array running at 50 MB/s, about 3.2 days. The rebuild reads every surviving drive: 28 TB in total.
Read errors
Reading 28 TB, the chance of at least one unreadable sector is 89% with desktop drives (1 per 1014 bits) and 20% 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 14 TB drives we would pick RAID 6. 4 drives in RAID 6 give the same 28 TB and survive two failures, or a read error during a rebuild.