OWC Jupiter Callisto, Быстрый старт, Страница: 17 / 30

OWC JUPITER CALLISTO STORAGE CONFIGURATION
14
RAID. Each member disk beyond the first will contain an exact copy of the information stored on the first disk.
• Stripe —
This is equivalent to RAID 0.
• RAIDZ1 —
This is equivalent to RAID 5. It is a stripe with one parity disk.
• RAIDZ2 —
This is equivalent to RAID 6. It is a stripe with two parity disks.
• RAIDZ3 —
This is a stripe with three parity disks.
4.2.3 Optimal RAIDZ Configurations
Due to the interaction between the 128K stripe size of RAIDZ and the 4K block sizes of modern high-capacity hard drives,
there is an optimal number of disks that each RAIDZ level (RAIDZ, RAIDZ2, RAIDZ3) should have for best performance.
When creating a new RAIDZ set, use the following formula to determine the optimal number of member disks:
2
n
+p
Where
n
is any number beginning at 1, and
p
is the number of parity disks in your RAIDZ level (RAIDZ=1, RAIDZ2=2,
RAIDZ3=3). Based on this formula, some of the optimal numbers of disks for each RAIDZ level are:
• RAIDZ1 —
3, 5, 9
• RAIDZ2 —
4, 6, 10
• RAIDZ3 —
5, 7, 11
Additionally, for optimal performance it is recommended to create a pool using more than one RAIDZ set. Write
performance especially will be negatively affected if only one RAIDZ set is used.
4.2.4 Creating Datasets
After the pool is created, the next step is to create devices from the pool. Think of the pool as one enormous hard disk
that you will be partitioning into multiple volumes. ZFS calls these volumes ‘datasets’, and there are two basic types of
datasets: filesystems and zvols. See
Section 4.1, ZFS Terminology
for more information on the types of datasets. If you
want to set up a NAS or use iSCSI to provide file-level data storage, you will create one or more filesystems. If you want to
use iSCSI to provide block-level data storage, you will create one or more zvols.
4.2.5 Creating a Filesystem
Filesystems are the most common datasets in the ZFS storage architecture. You can create a filesystem directly from the
pool, and you can also nest filesystems inside each other. To create a filesystem, click on the pool or the filesystem that
you want to use as the parent, then at the bottom of the screen click on the ‘Create ZFS Filesystem’ button to bring up
the following window (
Figure 4.2.5
).
Give your filesystem a name, then create it by clicking the ‘Add Filesystem’ button. There are advanced settings that
you can modify, including compression, deduplication, quotas, and reserved space, but if you are unfamiliar with these
options, it is best to leave them at their defaults.
Figure 4.2.5