Skip to content
IDX Cloud is live - deploy from Git in 30 secondsFree managed migration on every annual plan24/7 SupportIDX Cloud is live - deploy from Git in 30 secondsFree managed migration on every annual plan24/7 SupportIDX Cloud is live - deploy from Git in 30 secondsFree managed migration on every annual plan24/7 Support

RAID 0, 1, 5, 6 and 10 explained, and why RAID is not a backup

RAID protects a server from a failed disk. Choosing the right level is a trade-off between capacity, speed and how many drives can fail. It does not protect against deletion, corruption or ransomware.

IDX·October 19, 2026·2 min read

RAID combines several drives so that a server keeps working when one fails, or so that it reads and writes faster, or both. When you specify a dedicated server, choosing a RAID level is one of the most important storage decisions.

The common levels

  • RAID 0 (striping): data is split across drives for speed and full capacity. No redundancy: one failed drive loses everything. Only for temporary data you can rebuild.
  • RAID 1 (mirroring): two drives hold identical copies. Survives one drive failure; usable capacity is half. Simple and reliable, popular for operating system disks.
  • RAID 5 (striping with parity): needs at least three drives and uses one drive's worth of capacity for parity. Survives one failure. Rebuilding a large drive takes a long time, and a second failure during the rebuild loses the array.
  • RAID 6 (double parity): needs at least four drives, uses two drives' worth for parity, and survives two failures. A better choice than RAID 5 for large HDD arrays.
  • RAID 10 (mirrored stripes): needs at least four drives; data is mirrored and then striped. Half the raw capacity, but fast and resilient. The usual choice for busy databases.

How to choose

  • Operating system and small servers: RAID 1.
  • Databases and write-heavy applications: RAID 10.
  • Large capacity on HDDs for backups or archives: RAID 6.
  • Fast scratch space where loss is acceptable: RAID 0.

Hardware RAID controllers with battery- or flash-backed cache, and software RAID such as Linux mdadm or ZFS, can both be good choices; the important part is monitoring.

RAID is not a backup

RAID copies every change to all drives instantly. That includes a mistaken deletion, a corrupted file, an application bug and files encrypted by ransomware. RAID keeps a server running through hardware failure; backups let you recover from everything else. You need both.

Monitor the array

A degraded array that nobody notices is running without protection. Alert on drive failures and on early warning signs from drive health data, and replace failed drives promptly.

How IDX helps

The IDX dedicated server request form lets you choose drive types and counts and a RAID level, or ask us to advise. Our managed services include array monitoring, and our Data Backup service covers what RAID cannot.

IDX on digital infrastructure

Back to blog →

More in Storage

Storage

The 3-2-1 backup rule, and why modern threats added a 1 and a 0

Three copies, two kinds of media, one off-site. The rule has protected data for decades. Ransomware and silent failures have extended it to 3-2-1-1-0.

Storage

RPO and RTO explained: the two numbers that define your disaster recovery

How much data can you afford to lose, and how long can you be down? Agreeing on these two numbers turns disaster recovery from a vague worry into a design you can budget for.

Storage

Writing a disaster recovery plan people can actually follow

A disaster recovery plan is read under pressure, often at night, by whoever is available. It should be short, specific and tested.

Ready to move to hosting that keeps up?

Start today with a free migration and a money-back guarantee.

Get startedTalk to an expert