One of the upsides of being a paranoid bastard is that I keep regular backups, and often system notes. I’ll typically maintain a digital notebook recording the major software setup and configuration, mainly to track “What’s different about this one” and to assist in restoring should it become a manual affair.
Well, as it turned out, I recently had the opportunity for this as part of migrating my host environment to a newer OS. It’s been on my backlog for a while, waiting for a few kinks to be ironed out. So long in fact, that I said screw it, and just did a full restore the long way. In the process, I also discovered that my provider’s billing for backups and snapshots are pretty cheap, so I opted to take a snapshot in case things went sideways 🤣.
Amazingly, not only was it a relatively painless restore, but it also proved that a “real” operational restore from my nightly backups actually works in the production environment. These at best, have only been tested for offline cases not the live site, which for me is almost a cardinal sin, but a practical one.
Of course the part that was the real wrinkle? It wasn’t the nightly backup, or the OS change, it was misplacing or having failed to keep a proper system notebook. So rather than playing the “how the fuck do I …” game it was more an exercise in writing down the “that’s how the fuck it’s done.” Because it’s been a good while, and while care was taken to make the process largely obvious, it’s not good to not have a punchlist–so you know you didn’t bungle anything.
Yes, I’m paranoid like that (^_^)