zpool

status

• Test zfs installationd

list

zpool list

# all pools
zpool status

# specific pool
zpool status -L <pool>

create

• Create a storage pool

sudo zpool create \
  -f \
  -o ashift=12 \
  -O casesensitivity=insensitive \
  -O normalization=formD \
  -O compression=lz4 \
  -O atime=off \
  <pool> <vdev> <device-id1> <device-id2> <device-id3>

• -f: Force creating the pool. This is to overcome the "EFI label error".
• -m: mount point
• /<pool> by default on Linux
• /Volume/<pool> by default on macOS
• -d: disable all feature flags
• -o: Properties
• -O: File System Properties
• <pool>: arbitrary name for a pool
• <vdev>: type of the virtual device. E.g., raidz, mirror, etc
• mirror: requires at least 2 devices
• raidz1:
• <device-id>: disk identifier
• If an entire disk is used (e.g., /dev/sda), a GPT table is automatically created

set

• Set properties

# Compression
zfs set compression=lz4 <pool>
zfs set compression=on <pool>

export

• It's the equivalent of unmounting a device before unplugging

zpool export <pool>

# force if there are open applications using data
zpool export -f <pool>

import

• Scans the specified path and assemble an specified pool name
• Optionally pass an additional parameter to rename the existing pool

# by-id
zpool import -d /var/run/disk/by-id <pool>

# by-serial
zpool import -d /var/run/disk/by-serial <pool>

# by-path
zpool import -d /var/run/disk/by-path <pool>

# by-image-path
zpool import -d /var/run/disk/by-image-path <pool>

# by device name (not recommended as they vary)
sudo zpool import -d /dev <pool>

destroy

• Destroy any data and remove all metadata on a mounted storage pool

zpool destroy <pool>
zpool destroy <pool>/<dataset>

add

• Extends an existing pool

zpool add <pool> <device-id>

scrub

• Verifies the integrity of a pool

zpool scrub <pool>