RAID Levels Explained: RAID 0, 1, 5, 6 and 10
RAID (redundant array of independent disks) combines several drives into one volume for more speed, protection against a failed drive, or both. CompTIA A+ Core 1 (220-1201) objective 3.4 lists five levels: RAID 0, 1, 5, 6 and 10. This guide compares them, calculates usable space for any number of drives, and covers the RAID problems objective 5.2 asks you to troubleshoot.
The five RAID levels at a glance
| Level | Min drives | How it works | Usable space | Survives | Good for |
|---|---|---|---|---|---|
| RAID 0 | 2 | Striping | 100% of total | No drive failure | Speed and full capacity for data you can re-create, like a video scratch disk |
| RAID 1 | 2 | Mirroring | 50% with two drives | All but one drive | Two drives holding identical copies, such as a PC's boot drive |
| RAID 5 | 3 | Striping with distributed parity | Total minus one drive | One drive | File servers that need capacity and can lose one drive |
| RAID 6 | 4 | Striping with double parity | Total minus two drives | Two drives | Large arrays that must survive two failures, such as archives |
| RAID 10 | 4 | A stripe of mirrors (1+0) | 50% of total | One drive per mirrored pair | Databases and busy servers that need speed and redundancy |
RAID calculator
Assumes identical drives. With mixed sizes, most RAID controllers count every drive as the smallest one.
6 TB usable of 8 TB raw (75%)
Survives 1 drive failure.
How each level works
- RAID 0 (striping) splits data across every drive so they read and write in parallel. It is generally the fastest level, especially for writes, and keeps all the capacity, but it has no redundancy: if one drive fails, the whole array is lost.
- RAID 1 (mirroring) writes the same data to every drive. Two 2 TB drives give 2 TB usable, and the array keeps running as long as one drive survives.
- RAID 5 (striping with distributed parity) stripes data and spreads parity information across all the drives. Parity costs one drive's worth of space, so four 2 TB drives give 6 TB usable. It survives one failed drive; a second failure before the rebuild finishes loses the array.
- RAID 6 (striping with double parity) keeps two sets of parity, costing two drives' worth of space. It survives any two failed drives, which matters on large arrays where a rebuild takes hours.
- RAID 10 (a stripe of mirrors, also written 1+0) pairs the drives into mirrors and stripes data across the pairs. It keeps half the raw space and needs at least four drives, in even numbers. It survives one failure for certain, and more if each failure hits a different pair.
How to choose a level for an A+ scenario
Exam questions describe a need and ask for the level that meets it. Match the requirement to the trade-off:
- Speed and full capacity, and the data can be re-created or is backed up elsewhere: RAID 0.
- Two drives, and the system must keep running if one fails: RAID 1.
- Three or more drives, the most usable space of the redundant levels, and surviving one failure is enough: RAID 5.
- The array must survive any two drive failures at once: RAID 6.
- Four or more drives, and speed plus redundancy matter more than capacity: RAID 10.
Practice exactly this in the free Choose the Right RAID Level PBQ.
Troubleshooting RAID (objective 5.2)
Core 1 objective 5.2 asks you to troubleshoot drive and RAID issues. The symptoms it lists that point at the array:
- Array missing: the RAID volume no longer appears, for example after a controller or BIOS/UEFI setting changes, or when more drives have failed than the level survives.
- Audible alarms: many RAID controllers beep when the array is degraded, meaning a member drive has failed or dropped out.
- RAID failure and missing drives in the OS: check the controller's utility to see which member failed before you replace anything.
- Drive symptoms count too: LED status indicators, grinding noises, clicking sounds, bootable device not found, data loss or corruption, S.M.A.R.T. failures, extended read/write times and low IOPS.
When a member of a redundant array fails, replace that drive and let the array rebuild before anything else fails. In the A+ course, the RAID labs have you build an array in a RAID utility and diagnose a failing member.
Levels A+ leaves out
RAID 2, 3 and 4 are rarely used today, and nested levels such as RAID 50 and RAID 60 (stripes of RAID 5 or RAID 6 sets) belong to enterprise storage. None of them is in the 220-1201 objectives, so learn the five above for the exam. For the rest of Core 1's storage facts, see the A+ cheat sheet.