I’ve been running my node since last year with no issues, using all recommended hardware, SSD, name brand SATA cable, RPi power supply, etc. It has been solid, with 100% uptime between version reboots. Then out of the blue, I found my node not accessible, no ping/ssh/web/etc. Hard power down/reboot was the only way to recover access for less than 24 hours, when it would become unreachable again.
Digging through the dmesg.log file, I noticed a bunch of random virtual Ethernet ports and DHCP IPv6 errors, followed by “entered blocking state” and “entered forwarding state”, and eventually ending with some variation of the message “IPv6: ADDRCONF(NETDEV_CHANGE): veth2ece9ed: link becomes ready”.
Going back to my basic troubleshooting classes, I thought “What’s changed recently?” I realized I just installed a new router which has IPv6 enabled by default. I don’t use IPv6 on anything, so on a hunch, I disabled the IPv6 DHCP on the router. Now my node has been up and operational with no issues for the past week.
Not sure if this is just a fluke or some strange IPv6 bug with the Verizon FIOS router or Umbrel OS, but this seems to have solved my issues.