Hi everyone, I’m using TrueNAS
ElectricEel-24.10.2.3 as my home server. I installed Tailscale to access it remotely, but I can’t make it an exit node. Do I need to configure anything in addition to enabling "
Advertise Exit Node"? Thanks.
On the tailscale website you have to accept it in your network as an exit node
This machine does not expose any routes.
Is that from the tailscale website?
If you set the app to advertise exit node on Truenas, on the website under machines you should be able to click the menu for the machine and select edit route settings. The second section should be exit nodes where you should be able to turn it on as an exit node.
The exposing routes message seems like it is from the subnet routes section not the exit nodes section. The exit nodes section is below the subnet routes section.
Can you share what you see in the exit nodes section?
I also tried it on a second server I use as a test server, and the result is the same. I’m probably missing something.
Can anyone help me? I’m getting this message in the Tailscale dashboard. What can I try?
This device isn’t advertising itself as an exit node. Please run Tailscale up again with the --advertise-exit-node flag to enable this option.
Thanks, really.
Hi Luca,
I was frustrated byu the same issue.
It looks like a bug. The work arounbd is to open the command line for the app, and do the command:
tailscale set --advertise-exit-node
hen in the trailscale webpage console, the option to "use exit node"will no longer be greyed out.
See Advertise Exit node · Issue #2650 · truenas/apps · GitHub
This links to upstream issue: Docker: TS_EXTRA_ARGS doesn't respect --advertise-exit-node in the set sub-command · Issue #14496 · tailscale/tailscale · GitHub
Hope this helps - I have mine all working now with TrueNas as an exit node.
2 Likes
Mine works so I am not going to mess with it but I would think you could set the
advertise-exit-node in the Extra arguments or the Tailscale Daemon Arguments (not sure which right now, sorry) in the app setup.
For now I solved it by creating a second instance of tailscale in custom mode with the parameters:
General Settings
- Name:
tailscale-exit-node - Image Repository:
ghcr.io/tailscale/tailscale - Image Tag:
latest -
- Mount Path:
/var/lib/tailscale
- Mount Path:
- Volume Type:
Host Path - Host Path:
/mnt/pool/docker/tailscale-state - |
TS_AUTHKEY|tskey-auth-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx|
| — | — |
|TS_EXTRA_ARGS|--advertise-exit-node|
|TS_HOSTNAME|truenas-exit|
Hey,
i had the same Problem and my fix was “kinda” easy.
After adding the stuff mentioned in the Tailscale tutorials about the “Advanced Settings menu” and adding the Sysctl. settings i had to UNCHECK this
The “Run in Userspace” box. After that it worked like a charm but as it is not mentioned anywhere i didn’t dare to uncheck it.
sry… i tried to upload a screeshot 3 times but somehow the Forum didn’t like it.
If the exit node was already setup and working on an earlier version of Tailscale then unchecking Run in Userspace works. It is what I did on the system I have tailscale setup as an exit node to get it to work again as an exit node. That install was from before EE and is currently on Fangtooth, but continues to work as an exit node. The other system I have is current on Fangtooth has Tailscale setup, the Run in userspace not checked and if use as exit node is checked and the app stopped restarted it does not work and no exit node shows up in the Tailscale system for me to approve it as one.
you need to complete the robot welcome process under your forum account in order to post screen shots
1 Like
I am seeing the same issue. I never used my TrueNAS machine as an exit node before so I can’t speak to any different behavior in prior versions.
Reviewing the tailscaled commandline with ps aux inside the container, I see that checking the box doesn’t result in the --advertise-exit-node flag getting inserted. In my install (25.04), the “run in userspace” box was unchecked already by default.
Adding the flag in extra-args throws an error, which instructs the user to use the checkbox. Putting the flag in daemon-args is accepted into the config, but it gets stuck on deploying forever.
Curious if the exit-node feature is working or not for the majority of people…