TrueNAS ZFS Overhead: Why ZFS Uses More Space Than Expected
ZFS capacity is harder to predict than traditional RAID because several overhead mechanisms interact. The 3.1% slop space estimate used in Quick Mode is a minimum. Real overhead is typically higher.
See exactly how much usable storage your configuration delivers.
ZFS Slop Space (SPA Reserved)
The ZFS Storage Pool Allocator reserves 1/32 of total pool capacity (approximately 3.1%) as "slop space". This is a hard floor. ZFS will refuse to write data that would bring the pool below this threshold. The slop space prevents the filesystem from hitting completely full and becoming read-only.
This 3.1% is capped at 128 GiB for very large pools. A 200 TiB pool would have 128 GiB reserved, not 6.2 TiB.
Metaslab Alignment Overhead
ZFS divides pool space into metaslabs. Units of allocation management. The number and size of metaslabs is determined at pool creation and depends on vdev count and size. Poorly-aligned metaslab sizes can waste several percent of pool capacity in unusable fragments.
For most home NAS configurations, metaslab overhead is under 1%. For high drive-count pools with complex vdev layouts, it can be higher.
Record Size and Indirect Blocks
ZFS stores data in fixed-size records (default 128 KiB). Each record has an indirect block entry. For datasets with many small files, this overhead can be measurable. Though still typically under 1% for normal workloads.
Practical Capacity Planning for TrueNAS
For accurate TrueNAS capacity planning:
- Use the RAID calculator's Quick Mode for an initial estimate
- After pool creation, note the exact "Size" figure shown in TrueNAS's Dashboard
- Enter that value in the Exact Mode for a precise usable calculation
- Plan to keep the pool below 80% utilisation for stable ZFS performance
The jro.io OpenZFS Capacity Calculator is the most detailed ZFS capacity tool available and is recommended for complex vdev layouts.
Frequently Asked Questions
Why does TrueNAS show less space than expected even after accounting for RAIDZ?
Slop space, swap partitions (2 GiB per drive), and metaslab overhead all reduce the available capacity shown in TrueNAS's dashboard. The displayed "Available" figure in TrueNAS Storage already accounts for these deductions.
Does adding more vdevs change the ZFS overhead?
Each additional vdev has its own metaslab structure. Adding vdevs generally keeps overhead proportional, but complex pool topologies (mixing mirror and RAIDZ vdevs, for example) can interact in unexpected ways. Stick to uniform vdev configurations where possible.