A RAID 5 array is built from five disks of 4 TB each. About how much usable capacity does it provide?
A20 TB
B16 TB
C10 TB
D4 TB
RAID 5 uses one disk's worth of space for parity, spread across all disks. With N disks, usable capacity is (N - 1) times one disk: 4 x 4 TB = 16 TB. 20 TB is the raw total, which only RAID 0 delivers, and 10 TB is half the raw space, which is what keeping two copies of everything would give.
Question 2 Multiple Choice
A stripe holds data blocks D0 = 1011, D1 = 0110 and D2 = 1100, with parity P = D0 XOR D1 XOR D2 = 0001. The disk holding D1 fails. What value does the array rebuild for D1?
A0110
B1001
C0001
D1111
XOR is its own inverse, so D1 = D0 XOR D2 XOR P. 1011 XOR 1100 = 0111, and 0111 XOR 0001 = 0110, which matches the lost block.
Question 3 True / False
A mirrored RAID 1 array makes separate backups unnecessary.
TTrue
FFalse
Answer: False
Mirroring copies every write, including mistakes. An accidental deletion, file corruption or ransomware is copied to both disks immediately. Backups keep older versions in a separate place, which RAID does not do.
Question 4 Short Answer
RAID 4 keeps all parity on one dedicated disk, while RAID 5 spreads parity blocks across all disks. Why does RAID 5 handle many small random writes better?
Think about your answer, then reveal below.
Model answer: Every write must also update the stripe's parity. In RAID 4 every parity update goes to the same parity disk, so that disk becomes a bottleneck and small writes are serialized. RAID 5 rotates parity across the disks, so small writes to different stripes can update parity on different disks in parallel.
Each small write in RAID 4 or RAID 5 costs about four disk I/Os (read old data, read old parity, write new data, write new parity). Rotating the parity spreads that load instead of piling it onto one disk.