RAID-5 volumes are similar to striped volumes in that they stripe the data over multiple disk channels. In addition, RAID-5 volumes place a parity stripe across the volume. (Parity is a mathematical calculation performed on the data that provides information that can be used to rebuild data on failed drives.) If a single drive within the volume set fails, the parity information stored on the other drives can be used to rebuild the data on the failed drive. RAID-5 volumes require at least three physical drives (up to a maximum of 32 drives), using an equal amount of free space on all of the drives, as illustrated in Figure 4.5.
FIGURE 4.5 A RAID-5 volume set
FIGURE 4.5 A RAID-5 volume set
Parity Disk 1
Parity _ Disk 2
Parity Disk 3
RAID-5 Volume Set
Unlike with mirrored volumes, the system and boot partition cannot exist on a MING RAID-5 volume.
The main advantages of RAID-5 volumes are that they are fault tolerant and provide good performance because this configuration uses multiple disk I/O channels. The other advantage of RAID-5 volumes is that they require less disk space for fault tolerance than mirrored volumes need. A mirrored volume set uses half of the volume set to store the mirror. A RAID-5 volume set requires only the storage space of one drive in the volume set to store the parity information. For example, if you have three 100GB drives in a RAID-5 volume set, 100GB of the volume set is used to store parity information, and the remaining 200GB can store data. If your volume set contained five 100GB drives, you could use 400GB for data and 100GB for storing parity information.
The main disadvantage of a RAID-5 volume is that once a drive fails, system performance suffers until you rebuild the RAID-5 volume. This is because the parity information must be recalculated through memory to reconstruct the missing drive. If more than one drive fails, the RAID-5 volume becomes inaccessible. At that point, you must restore your data from your backup media.
As a best practice, mirrored and RAID-5 volumes should use identical hard drives (same manufacturer, model, and size) within the volume set. You should also purchase a spare drive that matches the drive set to facilitate recovery in the event of disk failure.
Real World Scenario
Selecting Fault Tolerance for Mission-Critical Servers
You are the network administrator of a large company. You need to install a Windows Server 2003 server that will be used as a database server. This is a mission-critical server and must be as reliable as possible. One of your tasks is to plan the disk configuration.
In this case you would most likely use a hardware implementation of RAID. When you purchase a high-end server, it is typically configured with drive arrays that can be configured to use RAID that is implemented as a hardware solution.
The difference between software RAID and hardware RAID is that software RAID is implemented solely through the software and requires no special hardware. Hardware implementations of RAID use special disk controllers and specific drives. When you install and configure the server, the manufacturer of the server typically provides software that is used to configure the RAID array.
Hardware implementations of RAID are more expensive to initially deploy than software implementations of RAID. However, RAID that is implemented through hardware is more fault-tolerant, faster, and is easier to recover from failure than the software RAID offered through Windows Server 2003.
For your disk configuration, you decide to buy a high-end server that supports RAID with hot-swappable drives, with an extra drive installed that can be used as an online spare. This allows you to have fault tolerance and speedy recovery in the event of disk failure.
Was this article helpful?