Why do raw drive capacity, RAID capacity and filesystem free space differ?
Raw drive sizes, array usable capacity and filesystem availability describe different layers. Redundancy, metadata, reserved space, snapshots and quotas can reduce or reshape what an application can use. ZFS manages pools and datasets together, so its reports should be interpreted through its own storage model rather than treated as ordinary md RAID output.
Before you start
- Authorised access to the dedicated server and the named inspection tools.
Instructions
1. Map ordinary block and filesystem layers
Compare the disk/volume tree with the actual application mount. Decimal drive labels and binary display units can also create apparent differences.
lsblk -o NAME,SIZE,TYPE,FSTYPE,MOUNTPOINTS
df -hT /srv/dataReplace /srv/data with the verified application path. Follow partitions, RAID and logical volumes before deciding where capacity is missing.
2. Use ZFS reporting for a ZFS deployment
Run these only where OpenZFS is installed and the intended pools are imported. They report state and dataset accounting without changing the pool.
zpool status
zfs list -o name,used,available,referenced,mountpointPool redundancy and dataset quotas/reservations/snapshots affect availability. A dataset view and a raw member-drive size are not interchangeable.
Check the result
- The capacity explanation identifies the exact layer responsible for the difference and any intended expansion procedure matches that layer.
Troubleshooting
- Adding a disk does not increase a dataset’s available space
- The storage design or quota may need a supported change; do not add or reshape a vdev without reviewing permanence and redundancy.
- df disagrees with a pool report
- Check dataset accounting, quotas and mount context before assuming a reporting error.
Official references
Use the documentation for your installed version when a flag, service name or platform feature differs.