My current setup involves a small Linux server that runs a few services (LoRaWAN server, some visualization dashboards, and VPN allowing external access to devices on that network). I want to make it more robust so there is no (or little) downtime in case the server breaks down.
Current idea: buy a second server with 100% identical hardware. I would put two partitions on it corresponding to two Linux installs. The "standby partition" would be always-on, every n minutes it sends a ping to the main server. If it's online, it synchronizes its root folder (rclone?) to the second partition so that it is always an exact clone of the main server. If it's down, after a grace period, it reboots under the "clone partition" so that services can resume ASAP.
Does this make sense? Is there a better way?
