Руководство Администратора для Cisco EFG250 EtherFast Network Attached Storage V 2.0
Configuring your Storage
Choosing a RAID Array Level
Linksys Small Business NSS4000 and NSS6000 Series Network Storage System Administration Guide
43
4
increases the reliability (for example, a two-disk RAID1 is half as likely to fail,
a three-disk RAID1 is one-third as likely to fail, and so on). Increased read
performance.
a three-disk RAID1 is one-third as likely to fail, and so on). Increased read
performance.
•
Disadvantages: Decreased aggregate storage capacity (each mirror disk
does not contribute to the total capacity of the array). Decreased write
performance. I/O intensive when resyncing mirrors.
does not contribute to the total capacity of the array). Decreased write
performance. I/O intensive when resyncing mirrors.
RAID5: RAID5 stripes data written to the array like RAID0, except that, unlike
RAID0, RAID5 generates parity information that also gets striped across the array.
This parity information is used during the reconstruction of the lost data when a
drive fails in the array. When a failed disk is then replaced, the array must
regenerate and rewrite the parity information of the array. This process is referred
to as “rebuilding”. During a rebuild, the array continues to be available for reads
and writes. If the parity information in the array gets out of sync with the data in the
array, usually as the result of an unexpected loss of power to the system, the array
must be rebuilt. Like RAID1, if a disk is missing from the array, the array is said to be
operating in “degraded” mode.
RAID0, RAID5 generates parity information that also gets striped across the array.
This parity information is used during the reconstruction of the lost data when a
drive fails in the array. When a failed disk is then replaced, the array must
regenerate and rewrite the parity information of the array. This process is referred
to as “rebuilding”. During a rebuild, the array continues to be available for reads
and writes. If the parity information in the array gets out of sync with the data in the
array, usually as the result of an unexpected loss of power to the system, the array
must be rebuilt. Like RAID1, if a disk is missing from the array, the array is said to be
operating in “degraded” mode.
•
Minimum Number of Disks: 3
•
Total capacity: m x (n-1)
•
Advantages: Increased reliability. The RAID array can sustain the loss of any
single disk without any data loss. Increased read and write performance.
single disk without any data loss. Increased read and write performance.
•
Disadvantages: Decreased aggregate storage capacity (you lose the
equivalent capacity of a single disk). CPU intensive calculation of parity,
especially during rebuild operation. I/O intensive during rebuild operation.
equivalent capacity of a single disk). CPU intensive calculation of parity,
especially during rebuild operation. I/O intensive during rebuild operation.
RAID10: RAID10 is a combination of RAID1 and RAID0. The NSS implements
RAID10 as two, 2-disk RAID1 mirrors, which are then used to build a RAID0 array.
RAID10 can be thought of as “mirror then stripe”.
RAID10 as two, 2-disk RAID1 mirrors, which are then used to build a RAID0 array.
RAID10 can be thought of as “mirror then stripe”.
•
Minimum Number of Disks: 4
•
Total capacity: 2 x m
•
Advantages: Increased read and write performance; increased reliability.
The RAID array can sustain the loss of any single disk without data loss.
Depending on which disks fail, the array can lose up to two disks without
data loss.
The RAID array can sustain the loss of any single disk without data loss.
Depending on which disks fail, the array can lose up to two disks without
data loss.
•
Disadvantages: Decreased aggregate storage capacity (you lose half of
the storage space).
the storage space).