The formula
RAID 5: capacity = (drives − 1) × capacity of each drive
Where it comes from
Every level trades one thing for another. RAID 0 stripes the data and uses a hundred per cent of it, but if one drive fails everything is gone: it is the only level that makes reliability worse than a single drive. RAID 1 writes the same data to two drives and leaves the capacity of one. RAID 5 uses the equivalent of one drive for parity and survives one failure; RAID 6 uses two and survives two. RAID 10 combines mirroring and striping: half the capacity, but the fastest rebuilds.
How to work it out by hand
- Multiply the drives by the capacity of each: that is what you bought
- Subtract the capacity the level gives up to redundancy
- RAID 5 loses one drive; RAID 6, two; RAID 10, half
- What remains is the usable capacity
What is worth knowing
One detail decides between RAID 5 and RAID 6 on large drives: the rebuild. When a 16 TB drive fails, rebuilding it means reading every other drive in full for many hours, and that is precisely the moment of heaviest load on drives that are usually the same age and from the same batch. A second failure during that window on RAID 5 takes the array with it. That is why past a certain size RAID 6 is recommended even at the cost of another drive. And it bears repeating: none of this is a backup. A file deleted by mistake disappears from every drive at once.