RAID combines multiple drives into one logical volume to gain speed, fault tolerance, or a blend of both. The level you pick decides how many drives you need, how much usable capacity you keep, and how many failures you can survive. Here are the five levels that cover almost every server build.
RAID 0 - striping (speed, no safety)
Data is striped across all drives, so every disk contributes to each read and write. You get maximum throughput and 100% of the raw capacity - but zero redundancy. One drive fails and the entire array is lost. Use it only for scratch data you can afford to lose. Minimum 2 drives.
RAID 1 - mirroring (simple redundancy)
Two drives hold identical copies. If one dies, the other carries on with no data loss. You keep only 50% of raw capacity, but rebuilds are simple and reads can be fast. Ideal for boot volumes and small, critical datasets. Minimum 2 drives.
RAID 5 - single parity (capacity plus one-drive safety)
Block-level striping with parity distributed across all drives. You lose one drive's worth of capacity to parity and can survive a single drive failure, rebuilding the lost data from parity. It is space-efficient and popular for general storage, but rebuilds on large drives are slow and stress the surviving disks. Minimum 3 drives.
RAID 6 - dual parity (survives two failures)
Like RAID 5 but with two independent parity blocks, so the array survives two simultaneous drive failures. That second parity is what makes it the safer choice for large, high-capacity arrays where a slow rebuild would otherwise leave you exposed. You give up two drives' worth of capacity. Minimum 4 drives.
RAID 10 - striped mirrors (speed and safety)
Mirrored pairs that are then striped together, combining RAID 1 redundancy with RAID 0 performance. It delivers excellent random read and write speed and fast rebuilds, and survives multiple failures as long as no single mirrored pair loses both members. The cost is 50% usable capacity. It is the go-to for databases and busy virtualization hosts. Minimum 4 drives.
Quick comparison
- RAID 0: max speed, no redundancy, 100% capacity, 2+ drives
- RAID 1: mirror, survives 1 failure, 50% capacity, 2 drives
- RAID 5: single parity, survives 1 failure, (n-1) capacity, 3+ drives
- RAID 6: dual parity, survives 2 failures, (n-2) capacity, 4+ drives
- RAID 10: speed and redundancy, 50% capacity, 4+ drives
One thing RAID is not
RAID protects against drive failure, not against deletion, corruption or disaster. It is not a backup. Keep a real backup regardless of the level you run, and consider a hot spare so the array starts rebuilding the moment a disk drops.
Bottom line
Scratch data - RAID 0. Boot and small critical volumes - RAID 1. General storage - RAID 5. Large arrays where rebuild risk matters - RAID 6. Databases and heavy I/O - RAID 10. And whatever you choose, back it up anyway.
Need the exact part?
Search our catalog of enterprise drives, memory, power and networking by exact OEM part number - so it seats and runs on the first boot.
Browse the catalog →