Help! Docker caddy suddenly no longer can find DNS records?

[rinkrealclose (Dan)]

@Whitestrake

If you can solve your Tailscale+VPN problem with a turn-key solution or in a dedicated VM, that significantly cuts down the amount of problem solving you have to do on your server.

While working on nextcloud/jellyfin I have removed the gluetun and its tailscale container portions so it shouldn’t be contributing to the problem. I am confused on the difference between gluetun and a virtual machine. It seems that a virtual machine would accomplish the same thing but add overhead. This person had it working: https://lemmy.world/post/7281194. I would be happy to use Mullvad, it’s just I already have a vpn purchased and don’t want to waste the subscription period while I have it.

@Whitestrake

I’d probably try something along the lines of docker compose exec caddy touch /media/server/custom/nextcloud/testfile and see if the testfile appears in the bind mounted directory or not. Sure fire way to tell if things are connected the way you think they are.

Wow, that totally didn’t make the testfile. I found this which suggested I do sudo chmod 755 /etc/caddyafter sudo chown -R root:root /etc/caddy. No dice there unfortunately, so I tried doing it via gui in dolphin admin mode

no testfile this way either. Is there another chown command I should be using, or a better file path that docker already allows? I did test touch from the host’s terminal and that created the testfile no problem.

@Whitestrake

But first, did you try renaming the Jellyfin compose service too? Did that also fail? Jellyfin should be a simpler set of circumstances to test, here.

Yes, I changed the caddy object to match jellyfin’s hostname, so both were jellyfin-ts instead of just jellyfin. Definitely caught up on the file permissions at the moment.