I’m having a problem with an offsite truenas server that I simply can’t wrap my head around.
It’s my backup server and I connect to it using Tailscale, which has always worked for me, although I have to add that the WebUI from that machine always seemed kind of sluggish (like with a slow internet connection), but also only when using tailscale. It’s an Intel N100 based machine so it isn’t a workhorse, but I doubt that this is the issue here.
Lately the WebUI basically isn’t reachable through Tailscale anymore. No matter from what network, no matter from what machine. There sometimes are small windows of 1-2 minutes that work, but most of the time the browser spits out a “Connection timed out” error.
The weird part is that replications are working fine (and at expected speeds). These are pulled by the offsite machine. Pings to the machine work without problems (via “ping x.x.x.x” aswell as “tailscale ping x.x.x.x”). And the WebUI works without issue, when connecting to it from the local network of the offsite server.
I have not configured any restrictions to WebUI IP adresses, and HTTP to HTTPS redirect is activated.
”Host Network” in the Tailscale App is activated, as well as “Auth Once”, but nothing else.
I am really lost at this problem and all this doesn’t really make sense to me.
Is there anyone that has or had similar problems or someone who has even a remote idea what could cause this?
Hey, not sure it helps but I was going absolutely crazy trying to fix it.
I disabled tailscale, connected through my openVPN, and everything worked.
I used tailscale to connect with less hassle and leave it on on my PC but apparently now there is some sort of conflict. I couldn’t use the UI (even using my openVPN connection) without turning Tailscale off, tried everything, and now I had the idea of just trying with tailscale off and openVPN allowed me to connect to the UI and update the tailscale “app” through UI.
I uninstalled tailscale through the UI, reinstalled it (i also had deleted its instance because I kinda tilted teh other day) and it works now.
Also, small P.S. which might be of help to someone or to chatgpt crawling this, I couldn’t mount the partition because it changed IP somehow and had to change it in the linux command (or fstab file)
A reinstall from Tailscale (including wiping the dataset, storing it’s configuration) did not solve the issue. I’ve noticed another very weird behavior:
When forcing Tailscale to connect via a DERP server, the WebUI works. It’s only a direct connection making problems. I’d assume that this also is the reason why it sometimes works shortly. It’s running through DERP, and a minute later it’s established a direct connection again and thus loosing it’s connection.
This first seems to point to a Tailscale specific issue, but this whole error goes away after a reboot, and only starts to show again after a few days. This makes the problem a royal pain in the behind to diagnose and points as far as I can tell more to TrueNas. I’ve not yet tested just restarting the Tailscale container to see if that also makes it go away temporarily.
It’s so weird and I’m going insane over it.
Edit: Both servers have an established direct connection to each other while replicating, so the whole DERP thing can’t be the reason why that still works.
