I'm fairly confident in our ability to recover from a disaster.
The 3 best pieces of advice I can give anyone in regards to this are:
1. Test your restores often. Never, for example, archive a yearly backup offsite without running several random test restores against files and directories from that backup.
2. Store full backups offsite. My yearly and monthly backups are sitting in a firebox at my home... away from the office.
3. And obviously, make sure your backup scheme is hitting all of the data you would need to recover and that you are getting the appropriate amount of historical snapshots for your needs. (nightly? weekly? monthly? snapshots etc)
For my system, I've got 4 offices with about 200GB of total "full" backup data and I run 1 tape backup server at each office with Symantec 11d in a classic
GFS system.
I've never had a total disaster, but I have had lots of partial restores to do over the years. I'd say they have been about 80% successful. When they have failed, it was always human error, and not machine or system error. Ie: The user was storing the data somewhere where it wasn't getting backed up and thus wasn't available for restore, etc.
Truth be told, I'm starting to trim down my methods. I have found them to be overkill for our needs. I am also exploring alternatives to tape and have started performing my weekly and monthly full backups to portable USB drives via backupexec's backup2disk feature.
Next time a tape drive blows out I will most likely replace it with a few portable usb hard drives. And I also plan to explore online backup solutions such as those offered by Vaultlogix and Mozy in the coming year.