⚠ Do now
Free tool · RAID calculator
Choose a RAID level and enter your drives. You get the usable space, how many drives can fail, and what a rebuild would have to read.
4 × 4 TB in RAID 5 = 12 TB usable. Survives 1 drive failure.
If every drive performed exactly at its rated URE spec, the chance of at least one unreadable sector while 12 TB is read to rebuild one drive would be 62%. With RAID 5, nothing is left to cover an unreadable sector during the rebuild, so the data in that stripe cannot be rebuilt, and many controllers stop the rebuild altogether.
The spec is a conservative bound, not a forecast, and real drives usually do better. The larger real risk is a second drive failing during a long rebuild, or a weak drive that has been quietly collecting bad sectors. If an array is already degraded and the data matters, have the drives imaged before the rebuild.
Decimal terabytes, as printed on drive labels (1 TB is 10¹² bytes); TiB is closer to what most operating systems show. RAID-Z figures ignore padding, metadata and reserved space. SHR figures are an estimate: Synology's RAID calculator is authoritative. Rebuild times assume the array is otherwise idle. What is a URE?
Array already degraded? Read its status first → Rebuild or stop? checker →
At a glance
| Level | Fewest drives | Drives that can fail | Usable space |
|---|---|---|---|
| RAID 0 | 2 | 0 | Every drive |
| RAID 1 | 2 | All but one | One drive |
| RAID 5 | 3 | 1 | All drives minus one |
| RAID 6 | 4 | 2 | All drives minus two |
| RAID 10 | 4, even | 1, up to one per mirror | Half the drives |
| RAID 50 | 6 | 1, up to one per group | All drives minus one per group |
| RAID 60 | 8 | 2, up to two per group | All drives minus two per group |
| SHR-1 | 2 | 1 | Total minus the largest drive |
| SHR-2 | 4 | 2 | Total minus the two largest drives |
| RAID-Z1, Z2, Z3 | 3, 4, 5 | 1, 2, 3 | All drives minus one, two or three |
| JBOD | 2 | 0 | Every drive |
Space. Sizes are decimal terabytes, as printed on the drive label (1 TB is 10¹² bytes). TiB, shown under each figure, is closer to what most operating systems report, which is why a NAS shows less than the label adds up to. RAID-Z figures ignore padding, metadata and reserved space. SHR figures are an estimate; Synology’s own calculator is authoritative.
Drives that can fail. The guaranteed figure is the worst case. RAID 10, 50 and 60 can survive more when the failures land in different mirrors or groups, but you cannot choose which drive fails.
Rebuild. The time assumes the array is otherwise idle at the rebuild speed you enter. The read-error chance uses the URE rating on the drive’s datasheet, the rated number of bits read per unrecoverable read error, applied to everything a rebuild has to read from the remaining drives.
It cannot see your drives. The URE rating is a conservative bound, not a forecast, and real drives usually do better. The larger real risk is a second drive failing during a long rebuild, or a weak drive that has been quietly collecting bad sectors.
If an array is already degraded and the data matters, have the drives imaged before the rebuild.
Questions
One. If a second drive fails before the rebuild finishes, the array goes offline and the data has to be recovered from the drives in a lab.
At least 3 for RAID 5, 4 for RAID 6, and 4 for RAID 10, which also needs an even number because the drives are paired into mirrors.
All the drives minus one. Four 8 TB drives in RAID 5 give 24 TB usable, about 21.8 TiB in your operating system.
Drives are sold in decimal terabytes, while most systems count in binary TiB, and the file system and the NAS keep some space for themselves.
Only if every file on it is also in a backup you have checked. A rebuild reads every remaining drive end to end. If the data matters, read the array’s status first with the RAID evidence reader, or call us before anything is rebuilt.
Checked 8 November 2026.
A drive has already dropped out
If the data matters, shut the array down cleanly and call. Our lab images every drive before any rebuild, then works from the copies.