Post #4455016

That unsettling feeling when you are not at home and your self-hosted services aren't available anymore. After checking out the infrastructure, I discovered that my whole rack went down and everything in it rebooted: Fileserver, auxiliary smart home nodes, core switch and core router. But all of it just rebooted, it wasn't off for long periods of time. It cannot be the fuse or RCD, these wouldn't reset themselves, so it must be something else. I restarted all my services and started digging for the cause. First hunch: the UPS. I know that the UPS battery is pretty much fucked, but I thought I could get by just a bit longer before ordering a new one. After checking the manual, I found out that this particular Eaton UPS does a battery test every week. Hmm... So I checked my power usage metrics for the rack on my Grafana. Turns out every week at the exact time when the devices rebooted last night, the current on the outlet drops from 1A to ~110mA. This must be the battery test. So my current hunch: The battery is so fucked, that the testing procedure makes the line voltage drop so far that everything rebooted. Time to order a new one, I guess. And I can sleep again now, knowing that my house won't burn down in the next few days until I'm back home. Also I should hook up that USB port of the UPS to at least one of my nodes :D