ZFS RAIDZ2
ZFS RAIDZ2
The ZFS sweet spot. Double parity, checksums, snapshots, used-block resilver. Strongly recommended for 4-10 drive pools.
How it works
RAIDZ1 plus a second parity (similar to RAID 6 but ZFS-native). Tolerates 2 simultaneous drive failures AND a URE during rebuild. Combine that with checksums detecting silent corruption and you have the strongest free RAID family for serious data.
Formula: (N − 2) × min(drives)
Pros / Cons
Pros
- Tolerates 2 simultaneous drive failures
- Tolerates URE during rebuild
- Full ZFS integrity stack (checksums, scrubbing, snapshots)
- Sweet spot for 6-8 drive pools
- Free, open source, mature
Cons
- Loses 2 drives of capacity
- ZFS RAM requirements (ECC recommended)
- Resilver still time-consuming for big pools
- vdev expansion is awkward (use stripe-of-mirrors for flexibility)
- Learning curve
When to use
Any serious ZFS pool from 4-10 drives. The default for TrueNAS, Proxmox, HexOS deployments.
When NOT to use
2-3 drive pools (use mirrors / stripe-of-mirrors). Pools 10+ drives may want RAIDZ3 for safety.
Rebuild math example
6 × 8 TB RAIDZ2 at 80% fill: resilver reads (6-2) × 8 × 0.8 = ~25 TB at 70 MB/s ≈ 33 hours. URE during resilver is tolerated thanks to second parity — array survives even with NAS-class URE rates.