This is just a quick rant regarding people who insist on relying on RAID as their sole backup solution.
Yes, RAID provides a certain amount of protection via redundancy, which may help save you from a drive failure, but does not constitute a backup solution. Usually the RAID setup you have isn’t fool proof – it may not be able to recover a damaged array and you’ll be mildly bitter that you have no real backups.
Besides, relying on a backup that’s based in the same location as the stuff you’re backing up is asking for a disaster. What if the server overheated and effectively destroyed all disks in the server? Or some clumbsy technician dropped your server upon moving it? Or even if the datacenter had a fire! You should always have a remote backup.
Another small point: I believe firmly that you should backup at least a weeks worth of data at a time, effectively allowing you to roll back any significant changes / issues that you notice after the fact (yes for the nerds, subversion and all of that fun stuff come in handy). I’ve witnessed far too many webmasters who have gotten their websites defaced to only have their “backup” solution backup the defaced website. Whoops?
And for the record, Hawk Host currently has two layers of protection – redundancy and backups. Not only do we run RAID-10 but we also perform full backups of each server. This includes the full partition so if needed we could restore the server from scratch (hopefully we don’t have to, yeesh). With these two layers of protection in place, we’ve yet to have any major data loss – and we hope to keep it that way.
Any who, sorry for the lack of effort put into this post for details, or coherency – it was just on my mind.