Tailscale version 1.14.3
Fedora 34
Inside my Tailnet I have a Synology NAS, an Android phone, a ThinkPad X1 Nano running Fedora 34, and an Always Free Oracle Cloud instance I use as my exit node. MagicDNS is on and looks like this:
The OCI instance is set up with sudo tailscale up --advertise-exit-node --accept-routes --reset
. It has a Tailnet IP of 100.97.186.66.
From my laptop I use sudo tailscale down && sudo tailscale up --reset --exit-node=100.97.186.66 --accept-routes
to connect.
The Synology NAS is acting as a subnet router to my local LAN with sudo /volume1/@appstore/Tailscale/bin/tailscale up --advertise-routes=192.168.99.0/24 --reset
. I can confirm the issue exists with or without the --advertise-routes
flag.
I am encountering some very odd behavior around certain services I use. The first is when attempting to install package from the Nixpkgs repository using, for example, nix-env -iA nixpkgs.bpytop
, it can’t resolve DNS:
querying info about '/nix/store/zz9y0ijs1kxr02xrr0yr1rp1ncmja4m1-bpytop-1.0.67' on 'https://cache.nixos.org'...
downloading 'https://cache.nixos.org/zz9y0ijs1kxr02xrr0yr1rp1ncmja4m1.narinfo'...
warning: unable to download 'https://cache.nixos.org/zz9y0ijs1kxr02xrr0yr1rp1ncmja4m1.narinfo': Couldn't resolve host name (6); retrying in 314 ms
The second is that if I use https://app.element.io to connect and sync with my Matrix homeserver, it works but any client on my machine will fail, both from the same laptop. That’s really puzzling to me.
Here’s my machine list (sorry for the imgur link but new users can only embed one image):
cat /etc/resolv.conf
on the laptop:
nameserver 192.168.99.1
nameserver 192.168.99.1
nameserver 192.168.130.11
search .
How do I debug this?