I know that ZFS reserves 1.6% disk space for copy-on-write. However, I am losing more than that. I have a ubuntu server with native ubuntu-zfs installed. The server has 45 drives, each 3TB. I am making three raidz2 arrays with 15 drives each:
#zpool create data raidz2 /dev/sd[b-p] raidz2 /dev/sd[q-z] /dev/sda[a-e] raidz2 /dev/sda[f-t] # zpool list NAME SIZE ALLOC FREE CAP DEDUP HEALTH ALTROOT data 122T 1.52M 122T 0% 1.00x ONLINE - # zfs list NAME USED AVAIL REFER MOUNTPOINT data 1.25M 98.7T 335K /data
3000000000000 bytes * 15 * 3 = 122.78 TB. So
zpool list output is correct.
3000000000000 bytes * (15 - 2) * 3 = 106.41 TB.
106.41 TB * (1 - 0.016) = 104.70 TB.
The available disk space should be
(1 - 0.016) * ((3000000000000 bytes * (15-2)) * 3) = 104.7 TB.
Why is ZFS only showing 98.7T?