EDACerton

MagicDNS and SSL are two different things. If you're getting an SSL protocol error on :5001, that likely means that whatever is running there is HTTP, not HTTPS.

Please provide diagnostics from within the plugin settings.

This is by design when an exit node is in use -- all traffic is routed to the exit node. If you want to be able to access local resources (including Docker containers) when an exit node is in use, you also need to set this: tailscale set --exit-node-allow-lan-access

You can probably get a similar effect by restarting the plugin (there's a button in the settings). The workaround that I posted will work in any version of the plugin. I won't make that automatic for the versions with the kernel bug because the kernel bug might not affect everyone, and automatically disabling Unraid services on the Tailscale IP would remove most of the reason folks use the plugin. The thing that strikes me as odd in your diagnostics is that something is restarting the Tailscale services a lot. Do you have a user script or something that could be forcing restarts of either Tailscale or the array?

Nothing changed related to SSH. Please post diagnostics from the Tailscale settings.

Soon There's some more work to be done before it's automatic, this is just getting ready for it In the meantime, the script will work just fine.

2024.08.28 This update contains an important alert for Unraid Connect users. We recently determined that the Flash Backup feature of Unraid Connect would back up the Tailscale state file. However, since Flash Backup is unencrypted, the state file should not be backed up. This update prevents the state file from being backed up, but existing backups will not be automatically erased. To remove any previous backups of the state file, do the following after updating the Tailscale plugin to version 2024.08.28 or later: On your Unraid server, go to Settings -> Management Access. Under Unraid Connect, deactivate flash backup. In the popup, select the option to also delete cloud backup. Reactivate flash backup.

I see that you have the "Use Tailscale Subnets" setting turned on in the Tailscale settings. I would turn that off; it's very rarely needed, and can cause weird connectivity problems in some networks. Aside from that, the plugin itself is detecting that the WebGUI is listening on Tailscale as expected, and from the diagnostics it looks like everything applied correctly:

I agree with your guess -- that kernel bug seems to be the issue. The bug was introduced in kernel version 6.1.103 (used by Unraid 6.12.12) and is still in the kernel in Unraid 6.12.13. If you roll back to 6.12.11, you shouldn't run into this. The Unraid 7 betas are also running an unaffected kernel version.

I'm not certain what happened initially, I would have to guess that somehow the Tailscale plugin got half-installed somehow during your recovery efforts. FYI -- Unraid reinstalls itself on every boot, so if you want to erase Tailscale, all you really need to do is remove: /boot/config/plugins/tailscale.plg /boot/config/plugins/tailscale/ And reboot -- it will be as if Tailscale was never there. (This is true for any plugin, actually.)

smb://server.local will only work on the local network. My general approach is to just use the Tailscale IP / Tailnet DNS (server.tailnet.ts.net) for any connections. That makes sure that things will work whether I'm local or remote.

Here is someone who did #3. You have to install a Tailscale docker container (this can be done in addition to the Tailscale plugin too if you're using the container from my repository):

I would start by turning off "Use Tailscale Subnets" and "Use Tailscale DNS". They are rarely required and can cause issues in some cases (e.g., I'd guess that DNS is your problem with the app store). I would also have IP forwarding turned on if you're doing subnet routing (this isn't a super big deal -- Unraid turns on IPv4 forwarding by default anyways, so the setting is more relevant for IPv6/exit nodes -- but it's a good idea to have the Tailscale setting turned on in case Unraid ever changes the default).

It looks like your system is having trouble communicating with the Tailscale control plane:

tailscale up

I just checked the Github side and it seems to be fine... What happens if you run this from the CLI? wget https://github.com/dkaser/unraid-tailscale-utils/releases/download/1.12.0/unraid-tailscale-utils-1.12.0-noarch-1.txz

This depends on how your containers are configured. If your containers are configured on bridge networks (i.e., you access them using the same IP that you access the Unraid WebGUI/SSH from), you can use the Tailscale plugin to connect to the containers in the same way (http://your.unraid.tailscale.ip:containerport/). If your containers are on macvlan/ipvlan networks (e.g., br0) where they get a different IP, then you won't be able to connect to them via the server's Tailscale IP. There are several ways to connect via Tailscale: Configure the Tailscale plugin to act as a subnet router. For linuxserver.io containers, use the Tailscale docker mod to install Tailscale inside the container itself. For other containers, use a separate Tailscale container, and use container networking to "sidecar" the Tailscale container to the service that you want to make accessible.

There are a number of different ways to do this. By default, Tailscale just passes network traffic -- if the container is serving HTTP, HTTP is what you will get from other endpoints. If you want HTTPS, there are a few different ways that you can do this: Configure the containers to serve HTTPS instead of HTTP (if possible). Use a reverse proxy like Traefik/SWAG/NPM to create an HTTPS endpoint. Use the reverse proxy built into Tailscale ("tailscale serve").

It looks like the configuration applied correctly, but for some reason Samba (the service that handles these file shares) didn't pick up the change. It's possible that there was something else going on at the time that prevented it from restarting (open files/file transfer/etc.). You can try using the Restart button inside the plugin settings; that will restart the services and might be enough to get it to work. Otherwise, I'd try restarting the server.

I'm not sure what to make of that... there could be something weird happening with how the Tailscale logs get redirected to the log file. Can you generate a diagnostic pack from inside the plugin settings for me?

For additional entertainment, what happens when @SpaceInvaderOne does a video for something? 😀

The major problem with your technique is that, by bypassing the GUI and making changes directly in the OS, you're creating several issues: The changes that you're making will be erased on every reboot. Changes to /etc aren't persistent in Unraid. The Unraid GUI isn't aware of the changes that you made in the OS (since the changes aren't in the "right" place), therefore it shows that the AD connection hasn't been established. I'd back up a few steps... restart the server, then try to join AD using the GUI. Assuming that fails, generate diagnostics / check syslog / check /var/log/samba and see if something gives a hint about why you're having trouble connecting.

Switch to advanced mode in the plugin settings, there's an option there to assign a port number.

Your diagnostics look fine. The WebGUI appears to be listening on the correct port: tcp 0 0 100.120.140.x:80 0.0.0.0:* LISTEN The plugin tests for a connection on the Tailscale IP as well, which is successful: 2024/08/05 18:36:18 rc.tailscale: Starting tailscaled: /usr/local/sbin/tailscaled -statedir /boot/config/plugins/tailscale/state -tun tailscale1 2024/08/05 18:36:18 tailscale-watcher.php: Starting tailscale-watcher 2024/08/05 18:36:33 tailscale-watcher.php: Tailscale IP detected, applying configuration 2024/08/05 18:36:33 tailscale-watcher.php: /usr/local/sbin/tailscale set --accept-routes=false 2024/08/05 18:36:33 tailscale-watcher.php: /usr/local/sbin/tailscale set --accept-dns=false 2024/08/05 18:36:33 tailscale-watcher.php: /usr/local/sbin/tailscale set --stateful-filtering=false 2024/08/05 18:36:33 tailscale-watcher.php: Restarting Unraid services 2024/08/05 18:36:48 tailscale-watcher.php: WebGUI listening on 100.120.140.x:80 2024/08/05 18:37:48 tailscale-watcher.php: WebGUI listening on 100.120.140.x:80 2024/08/05 18:38:48 tailscale-watcher.php: WebGUI listening on 100.120.140.x:80 2024/08/05 18:39:48 tailscale-watcher.php: WebGUI listening on 100.120.140.x:80 2024/08/05 18:40:48 tailscale-watcher.php: WebGUI listening on 100.120.140.x:80 2024/08/05 18:41:48 tailscale-watcher.php: WebGUI listening on 100.120.140.x:80 Based on the error message you're describing, I would check to see if there's something weird happening with your management device. Also, make sure that you're using the new Tailscale IP for the Unraid server; it will change after running "Erase Configuration". You could also try adding the Tailscale hello server to your tailnet, then try to access it from your phone/laptop just to make certain that everything there is working properly: https://tailscale.com/kb/1073/hello

Posting diagnostics is common on the forums, so I wouldn't be worried about it, but if you're concerned you can PM me the diagnostics file instead. (I will respond here though for the benefit of search/other folks in the future.)