Disk drives are mechanical devices with platters that typically spin at between 7,000 - 10,000rpm - over time it is inevitable that the disks in your server (or workstations) will eventually fail. With proper planning however, the consequences of these hardware failures can be minimised.

Disk drive failures are more complex to resolve than other hardware issues (e.g power supply, memory). For example, if the power supply in your server were to fail, while there would be some "down-time", once the power supply had been replaced the server would be ready to resume operation as before. In contrast, once a failed disk drive has been replaced, all the information that was previously stored on the original disk drive must be restored.

For this reason a single-disk volume on a server, is referred to as not being "fault-tolerant" or "redundant". A failure of the disk drive in a 'single-disk' configuration will result in a catastrophic "crash". Time to recover from such a failure will depend on the amount of data to be restored and the type of backups held. The best-case scenario is 1/2 day and if an "offline backup" is not held 2-3 days is not uncommon.

Given the amount of disruption that can be caused by the failure of a disk drive in a single-disk server, a fault-tolerant disk configuration is highly recommended. There are a number of options available depending upon capacity, budget and performance requirements. RAID1 (mirroring) and RAID5 (array) are popular options. In both cases, the failure of a single-disk within the server does not disrupt the operation of the server.

A fault-tolerant disk configuration on it's own though is not enough. Say there was a virus strike that could not be repaired or say someone was logged onto the server with administrator rights and inadvertently damaged the system volume preventing the server from being able to restart. There is no substitute for having good backups of the operating system and user data.

One way that backups can be classified is whether they're "on-line" or "off-line". An "on-line" backup is recommended for regular backup of user data and other important data on the server (like e-mail). While an "off-line" backup is recommended for backup of the server operating system and associated system files (registry, directory etc).