OWC JUPITER CALLISTO
STORAGE CONFIGURATION
12
STORAGE CONFIGURATION
4.1 ZFS TERMINOLOGY
In order to use Jupiter Callisto most effectively, it is important to have an understanding of the following terms and concepts
relating to the ZFS file system and volume manager.
Pool —
Also known as ‘ZFS pool’ or ‘zpool’. A collection (or ‘pool’) of all of the virtual devices (vdevs) that have been configured
for use. All of the vdevs inside a pool are automatically striped together. The pool is the largest logical construction in the
ZFS environment. Other ZFS components (filesystems, zvols, snapshots, and clones) can be created inside the pool. These
components are called ‘datasets’ and they are the closest things that ZFS has to what would be a ‘volume’ or a ‘partition’ in
a traditional volume manager.
Vdev —
A vdev, or virtual device, is a building block of the pool. There are two basic types of vdev: low-level and high-level.
A low-level vdev can be an entire hard drive or a partition on a hard drive. A high-level vdev is made up of multiple low-level
vdevs, and can be thought of as a RAID. In ZFS, a high-level vdev can be a stripe, a mirror, or one of three levels of ‘RAID-Z’.
Dataset —
This is a blanket term for any ZFS component created out of a pool. These are the devices that will be used to
store data and that can be manipulated by users. The types of datasets are: filesystems, zvols, snapshots, and clones.
Filesystem —
This term can be confusing, because it has a slightly different meaning in ZFS than you might expect. In ZFS,
a filesystem is a dataset that provides file-level access to data. It is similar to a folder, but with added capabilities. A filesystem
is required in order to create a NAS.
Zvol —
A ZFS volume (or zvol) is a dataset that provides block-level access to data. It behaves essentially like a hard drive as
far as an end user is concerned, and is required in order to set up a SAN.
Snapshot —
A read-only copy of a filesystem or zvol. Because of the copy-on-write transactional model that ZFS uses, a
snapshot can be created very quickly and takes up no additional space initially, because instead of a new copy of all of the
information stored on the source, a snapshot contains pointers to the original data.
Clone —
A writable copy of a filesystem or zvol. A clone can only be created from a snapshot, and like a snapshot it can be
created very quickly and initially takes up no additional space. As changes are made to the clone, the storage space that it
takes up will grow.