EDACerton

Please install plugin diagnostics from community applications and post Tailscale diagnostics.

The plan is to keep it as a plugin, for the specific reason that you stated

I would guess that your connection is being relayed, which will slow things down a lot. You can check that by running "tailscale status" from the Unraid CLI.

If you think there's a bug in the Tailscale DNS implementation, that should go to https://github.com/tailscale/tailscale .

MOTD displays system information (CPU, memory, disk usage, service status, etc.) after SSH/console login. Configuration MOTD can be customized by creating a configuration file on the flash drive. Details are available on Github.

Unless you have a specific reason to be setting --accept-routes and --accept-dns, you should turn both of those off. Those are rarely needed, and tend to cause problems if you haven't planned for using them. Also, make sure that you're approving the routes in the Tailscale admin console after you advertise them.

Please follow the instructions in the pinned post to generate diagnostics using "Plugin Diagnostics" from Community Applications.

Tailscale won't try to use port 443 by default. This sounds like you might have enabled serve or funnel in Tailscale. What is the output if you run tailscale serve status ?

The default setting is No: https://github.com/unraid/unraid-tailscale-utils/blob/d17e399f45b3c13a2e6b6f45f22f27d792efa817/src/usr/local/emhttp/plugins/tailscale/settings.json#L11

For most cases, it is best to have MagicDNS turned off, because it doesn’t provide any benefit (and can cause unnecessary issues, as folks are seeing). (SpaceInvaderOne did a great video on setting up Tailscale, but the one problem in it is that it suggests that you have to have MagicDNS turned on — you don’t need it, and probably shouldn’t turn it on. )

This should be fixed now... some unintended side effects from the plugin migrating to Unraid's official repository last night.

@Squid

I'd try turning subnet routes off. For the NAT-PMP, you'd have to make certain that it is properly configured on pfSense: https://docs.netgate.com/pfsense/en/latest/services/upnp.html

https://selfhosters.net/remote/tailscale/advanced/

It sounds like you're effectively creating two different "exit nodes".... one with Wireguard and one with Tailscale. If you want to use both, it might be better to connect the containers via WireGuard-Docker.

I can see that you're advertising a subnet route from your Unraid server. Is the setting to use Tailscale subnets enabled on her desktop? Otherwise, it's fairly difficult to read your log files... it seems like NAT-PMP on your router and Tailscale aren't playing well together, plus there's a lot of nginx messages in the syslog.

One interesting aside on this -- I recently started having a similar problem on an Asus Prime Z790-P Wifi board after I applied a BIOS update. I haven't had time to play with it much, but I've run into the "black screen" several times, with the server then booting after I open the BIOS settings.

I requested that you create a Tailscale diagnostics package following the instructions in the pinned post.

Sorry, guess I could have been more clear.... get rid of everything in the Samba extra configuration. It should be blank. (And it might be worth trying to connect after you do that... the Samba config has your Tailscale IP, I think it's those extra settings that are causing difficulty.)

Please do this: Remove the Samba extra configuration. Restart Tailscale using the button in the Tailscale settings. Follow the instructions in the pinned post to generate Tailscale diagnostics using Plugin Diagnostics.

Please follow the instructions in the pinned post in this topic to generate a Tailscale diagnostics package using the "Plugin Diagnostics" app.

FYI -- there is a new way to connect containers together coming in Unraid 7 beta 3 that would probably be an alternate solution for this (think NPM docker connected to Tailscale docker) that doesn't require the extra upkeep of LXC containers (running updates, etc.)

Reason #7,236,742,146,894,167,684 to hate printers I thought that it might have had something to do with tailscale1 still being in "extra interfaces" even though the interface was gone, but I just tested adding a nonexistent device and Unraid works fine (even after a reboot), so that doesn't seem like it would be the issue. Can you generate a diagnostics pack for the plugin? Instructions are in the pinned post.

There's nothing in your log that indicates any problem. Please create a Tailscale diagnostic package by following the instructions in the pinned post.

Uninstall plugin, restart server -- it will be like the plugin was never there. The plugin shouldn't affect the local IP, but I have a hunch that I want to look into later. However, one other note -- you don't have to uninstall the plugin if you need NetBIOS -- you can ignore that warning, it just means that you won't be able to access SMB shares via Tailscale. Other things (WebGUI, SSH, containers, etc.) will still work.