 | By Fred Langa Drive mirroring via RAID would appear to be a great way to keep an updated backup at the ready. Unfortunately, the many RAID “gotchas” make this system impractical for backing up most PCs. |
Mirroring copies the errors along with the data
Andy’s having trouble with his drive-mirroring setup:
- “I use an external drive system by D-Link — great little unit — that has two SATA drives. I have mine mirroring each other. If one fails, the other picks up straight away. This drive is Ethernet-networked and seems to work on some type of tiny, internal Linux operating system. Some applications do occasionally seem to have problems seeing files on the drives. Is this because the drives don’t use NTFS?”
Drive mirroring is a kind of continuous, real-time backup where an OS keeps a primary and a secondary drive in perfect sync. Whatever changes are made to one drive are transferred automatically to the other.
Mirroring is actually a form of RAID (Redundant Array of Inexpensive Disks). This ancient hard-drive technology originated way back in 1987, when MS-DOS 3.3 was king and hard drives were incredibly expensive. For example, a 10MB hard drive cost $900 in 1987 dollars. (And yes, that’s MB, not GB.)
Today, there are several major types of RAID arrays. (A good description of the varieties is in a tutorial by the Advanced Computer & Network Corp.) Some of these configurations are useful in specialized applications, such as always-on, mission-critical data centers. But for ordinary PCs, a simple RAID mirroring setup is a poor backup solution. Drives aren’t expensive any more. Right now, 10 megabytes of hard-drive storage typically cost less than a penny, not $900. You can inexpensively buy a second hard drive and use that for periodic backups of your primary drive in case an unrecoverable problem develops.
Related posts:
