What about checkpoint backups?
I've been curious about hard disk backup for quite a while and I found the posts here to be thought-provoking.
We periodically take "checkpoint" backups of our data, which go on tape and go "on the shelf" essentially indefinitely, so that in the event of a problem that is not immediately apparent, we have a "known state" of our data at a point in time.
George's calculations seem to ignore this backup need. If I have a set of tapes, the incremental cost of a checkpoint backup is the cost of media. But I gather that with hard disk backup, I need another set of drives, server, etc. I would like to take checkpoints at key times: month-end, quarterly, end-of-year, at-major-change-point. It seems that a disk solution becomes much more expensive then.
One might argue that all I need is a set of drives, but this requires me to install them in my backup server before they are usable. With a set of tapes, I load them in my autoloader and restoring right away. In an emergency situation, a clean step-by-step recovery procedure builds confidence in the eyes of my stakeholders. They would not want to have to start assembling hardware before we can recover.
I'm very interested in how those who rely on disk backup handle this situation. Thanks!
