Estimate usable capacity, speed gain, and fault tolerance for a RAID array
Disk performance (optional)
Results
Speed gain is shown relative to a single disk, based on typical random I/O write-penalty factors (RAID 0/1E/10 = 2, RAID 5/50/5E/5EE = 4, RAID 6/60 = 6). Sequential throughput on real hardware is often higher than the write figure shown here. RAID 50/60 assume the disks are split into 2 equal parity groups.
Choosing a RAID level
RAID 1 — two mirrored disks; simple and safe for a boot volume or a small server.
RAID 10 — the best fit for databases and virtual machines: fast random writes and quick rebuilds.
RAID 5 — good capacity, but avoid it with large hard disks (4 TB and up): rebuilding takes many hours, and a second failure or read error during the rebuild loses the array.
RAID 6 — survives two failed disks; the usual choice for large-capacity file and backup storage.
RAID 0 — speed only, no protection: one failed disk loses everything.
RAID is not a backup
RAID keeps a server running when a disk fails. It doesn’t protect against deleted files, ransomware, a failed controller or a fire. Keep separate backups — ideally the 3-2-1 rule: three copies, on two kinds of media, one off-site.
mrpanos.com