You’ve spent thousands of dollars on that cool technology; clustering, redundant controllers, redundant disks, redundant power supplies, redundant NIC cards, multiple network drops, fancy tape backup devices and the latest and greatest tape technology. You’re all set. There’s no way your going to have downtime.
But one day something does go wrong, is it the disks, no way you’ve implemented RAID with hot swappable hard drives, is it the server, can’t be you’ve got your servers clustered and any downtime due to failover would be so small that hardly anyone would even notice it. Well if it’s not your stuff it must be the network. Those guys are always making changes and not letting you know about it until there’s a problem. No, checked with them, no changes. What’s left? It must be the application. Once again that application group rolled out a new version and no one informed the DBAs. Once again foiled, the application team says no changes went out.