On Thu, 13 Nov 2008 19:38:31 -0800, "William Sommerwerck"
wrote:

Hmmm... I'm backing up to a drive that runs only when it's being backed up
to. So that, in theory, would pretty much eliminate the possibility of
simultaneous failure.

Only if it's a different maker, model, or lot number drive. Read my
rant again. One issue that I listed was having two identical drives,
one spinning, the other just sitting on the shelf, fail roughly
simultaneously. That can happen if the failure is caused by IC
package leakage, board contamination, tin whiskers, or divine
irritation. I only saw this problem once, so I suspect it's rather
unusual. Still, it's something to worry about. These daze, I avoid
running backups to identical drives out of paranoia.

Actually, what I'm currently doing for desktops is probably equally
dangerous. After about 4 years running, I buy a bigger|better|faster
drive, mirror the contents of the old drive to the new drive, and then
use the old drive for backups. In other words, pre-emptive
replacement. Using various mirroring software, it's quite quick and
easy. Convincing the customers to replace a working drive is not so
easy. The theory is that the new drive will be more reliable and last
longer than the older backup drive. So far, it's been working well.
However, I now have a growing pile of very used "backup" drives that
are unsuitable for new systems or replacements.

For the servers, I like to use an active backup server. Only the data
gets synchronized with rsync, rdist, rcp and similar software. If the
main server is having a bad day, users simply point to the backup
server and blunder onward. The backup server is usually in a
different part of the building or off-site (in case of fire). Some of
these also have a tape drive, which I use sparingly.


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558