How should I choose between RAID 0, 1, 5, 6 and 10?
Choose RAID according to required fault tolerance, usable capacity, write behaviour and recovery time. RAID 0 stripes without redundancy; a two-drive RAID 1 mirrors data; RAID 5 and RAID 6 use single and dual parity. RAID 10 combines mirroring and striping, with survivable failures depending on which mirror members fail.
Before you start
- Authorised access to the dedicated server and the named inspection tools.
Instructions
1. Identify the existing layout
On Linux software RAID, inspect the current array before designing a change. Hardware controllers require their own supported inventory tool.
cat /proc/mdstatThe displayed level and member state apply to Linux md arrays only. An empty report does not mean a hardware RAID controller has no arrays.
2. Evaluate the workload and recovery requirement
Compare random writes, sequential traffic, disk size, rebuild duration and degraded-state performance. Include controller cache protection and replacement-drive availability in the design. Do not calculate safety only from nominal usable terabytes.
Check the result
- The selected layout meets the intended drive-failure scenario and measured workload requirements, with a separate recovery strategy for deletion, corruption and complete server loss.
Troubleshooting
- RAID is being treated as a backup
- Redundancy can preserve service through some drive failures, but it also mirrors deletion or corruption and does not cover every failure domain.
- A new level is needed on an existing array
- Reshaping can be disruptive and implementation-specific. Plan a supported migration with verified recovery, not an improvised create command.
Official references
Use the documentation for your installed version when a flag, service name or platform feature differs.