Hi,
I’ve had a Home Assistant instance running for a few years inside a Proxmox VM and it’s been rock solid with no issues at all.
When I went to bed last night I noticed that the downstairs lights hadn’t dimmed like they usually do but didn’t think anything of it. Then this morning I started experiencing weird issues which seemed to be DNS related.
I have a local DNS Server that I use for routing to named hosts internally but the routing is configured so that first traffic is routed to the instance of AdGuard I have running in Home Assistant. From there it then forward onto my internal DNS before routing out to the internet.
It turned out that my Home Assistant instance had crashed with a memory error some time yesterday. It’s the first time i’ve ever seen such an issue but to get things running quickly I simply powered it off and then back on, and sure enough it fired up.
However, I still have a problem because it looks like AdGuard is having some problems setting up it’s DNS listening socket. I see this error in the logs:
[fatal] couldn't start forwarding DNS server: couldn't listen to UDP socket, cause: listen udp [fe80::f6d9:8e63:bdbc:f15b]:53: bind: invalid argument
I haven’t made any significant changes recently and so far as I can see in my list of installed components there is nothing running which should also be trying to grab that UDP port. I know typically you’d see that error if you were trying to run more than one DNS Server on the same host.
Does anyone have any suggestions or clues on how I can troubleshoot this one please?
For the time being i’ve simply changed the routing so I go direct via my other DNS Server (this is DNSMasq which is running in a totally seperate container well away from HA).
Thanks
Pete