4 × 6 TB in RAID 5: the numbers
| Usable space | 18 TB (16.4 TB in Windows) |
|---|---|
| Raw space | 24 TB |
| Used for protection | 6 TB |
| Efficiency | 75% |
| Drives that can fail | Any 1 |
| Rebuild at 150 MB/s | 11 h |
| Rebuild at 50 MB/s (busy) | 33 h |
| Data read during rebuild | 18 TB |
| Read error risk, desktop drives | 76% |
| Read error risk, NAS drives | 13% |
Rebuild
Replacing a failed drive means writing a full 6 TB drive. At 150 MB/s with no other load that takes about 11 h; on a busy array running at 50 MB/s, about 33 h. The rebuild reads every surviving drive: 18 TB in total.
Read errors
Reading 18 TB, the chance of at least one unreadable sector is 76% with desktop drives (1 per 1014 bits) and 13% 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 6 TB drives the rebuild is short enough that RAID 5 is a reasonable choice, as long as anything you cannot replace is also backed up elsewhere. RAID 10 with the same drives gives 12 TB, faster writes and shorter rebuilds.