A week in the fjords, and my self-hosted server survived
The cover photo is my favourite moment of the trip: a rocky ledge above a fjord in western Norway, no signal, and for once no urge to check it. I was offline for a week. The self-hosted server that runs our hackathon community was not, and it was fine. Past trips did not go this smoothly, so this time I prepared properly.
Backups you've actually restored
A backup you've never restored is just a hopeful file. Two weeks before the trip I restored the previous night's database and uploads to a spare VPS and clicked around. It took an evening and found one folder my script had been skipping for months.
Alerts that go to more than one person
Uptime and disk-space alerts now go to me and to one trusted co-admin. She had access, a one-page runbook and permission to restart services without asking me first.
No upgrades the week before
I froze everything: no package updates, no new modules, no "quick" config tweaks. Boring servers are happy servers.
Leave the disk room to breathe
Logs and media uploads keep growing while you're away. I cleared out old logs, set up rotation properly and made sure there was plenty of free space.
The result: one alert all week, a certificate renewal that sorted itself out, and zero messages from my co-admin. I spent the time hiking and eating far too many cinnamon buns instead. Highly recommend both.
Info
html
asdasdasd