Administrator Guide

data disks and the one disk providing parity is the parity disk. In reality, the parity is distributed among all the disks, but
conceiving of it in this way helps with the example.
Note that the number of data disks is a power of two (2, 4, and 8). The controller will use a 512-KB stripe unit size when
the data disks are a power of two. This results in a 4-MB page being evenly distributed across two stripes. This is ideal for
performance.
● Example 2: Consider a RAID-5 disk group with six disks. The equivalent of five disks now provide usable capacity. Assume
the controller again uses a stripe unit of 512-KB. When a 4-MB page is pushed to the disk group, one stripe will contain a
full page, but the controller must read old data and old parity from two of the disks in combination with the new data in
order to calculate new parity. This is known as a read-modify-write, and it's a performance killer with sequential workloads. In
essence, every page push to a disk group would result in a read-modify-write.
To mitigate this issue, the controllers use a stripe unit of 64-KB when a RAID-5 or RAID-6 disk group isn't created with a
power-of-two data disks. This results in many more full-stripe writes, but at the cost of many more I/O transactions per disk
to push the same 4-MB page.
The following table shows recommended disk counts for RAID-6 and RAID-5 disk groups. Each entry specifies the total number
of disks and the equivalent numbers of data and parity disks in the disk group. Note that parity is actually distributed among all
the disks.
Table 41. Recommended disk group sizes
RAID level Total disks Data disks (equivalent) Parity disks (equivalent)
RAID 6 4 2 2
6 4 2
10 8 2
RAID 5 3 2 1
5 4 1
9 8 1
To ensure best performance with sequential workloads and RAID-5 and RAID-6 disk groups, use a power-of-two data disks.
Disk groups in a pool
For better efficiency and performance, use similar disk groups in a pool.
● Disk count balance: For example, with 20 disks, it is better to have two 8+2 RAID-6 disk groups than one 10+2 RAID-6 disk
group and one 6+2 RAID-6 disk group.
● RAID balance: It is better to have two RAID-5 disk groups than one RAID-5 disk group and one RAID-6 disk group.
● In terms of the write rate, due to wide striping, tiers and pools are as slow as their slowest disk groups.
● All disks in a tier should be the same type. For example, use all 10K disks or all 15K disks in the Standard tier.
Create more small disk groups instead of fewer large disk groups.
● Each disk group has a write queue depth limit of 100. This means that in write-intensive applications this architecture will
sustain bigger queue depths within latency requirements.
● Using smaller disk groups will cost more raw capacity. For less performance-sensitive applications, such as archiving, bigger
disk groups are desirable.
Tier setup
In general, it is best to have two tiers instead of three tiers. The highest tier will nearly fill before using the lowest tier. The
highest tier must be 95% full before the controller will evict cold pages to a lower tier to make room for incoming writes.
Typically, you should use tiers with SSDs and 10K/15K disks, or tiers with SSDs and 7K disks. An exception may be if you need
to use both SSDs and faster spinning disks to hit a combination of price for performance, but you cannot hit your capacity needs
without the 7K disks; this should be rare.
176
Best practices