Apps on VLAN can’t reach each other

TrueNAS General
1

I bought my own domain and I am now trying to setup nginx-proxy-manager (npm) to access my truenas-apps like: jellyfin.example.com or truenas.example.com

I think I already got DNS set up correctly and added a Let’s Encrypt certificate in npm and I added the proxy hosts in npm.

To get that working I created a VLAN in Truenas itself to have the possibility to give an app its own IP. So I gave npm its own IP (10.0.30.11) so it still can run on the default ports, which is otherwise not possible except you would change either the Truenas-GUI-port or the npm-GUI-port.

For some reason it now seems like the apps can’t access each other, but can access the internet (tested with curl). Which I think is the reason why the access via npm (with the domains) does not work. From my client host I can access the services directly with their IP though.

  • Truenas-IP: 10.0.30.10
  • npm-IP: 10.0.30.11
  • Jellyfin-IP: 10.0.30.40
  • Pi-hole-IP: 10.0.30.10

image
image911×420 8.67 KB

  • br0 bridge-member: enp9s0
  • vlan30 parent: enp9s0
  • br30 bridge-member: vlan30

I don’t know pretty much anything about what I did with those interfaces. I kind of just tested what works and what now.
The br30 I created, because I could not add the IPs directly to the VLAN for some reason, without it failing.

I also added the 10.0.30.10 IP to br30 so my Pi-hole can use it, which probably isn’t good either, but it seems to work. The Truenas-GUI also only runs on 10.0.30.10.

2
  1. How did you exactly specify IP for each app?
  2. Can you access npm via 10.0.30.40?
  3. Post the output of ip a, please.
3

Okay, so I got it working now and it was totally my fault.
I reset the interfaces, created all of them again and restarted the system.

IP-Assignment:
Since the br30 contains some IP-Addresses I could just select one as host-address in the app-configuration. This was confirmed as working since I could access the apps afterwards over the selected IP.

By reconfiguring the apps, I also noticed that nginx-proxy-manager actually still ran on the default ports (http: 30021 & https: 30022). I changed those to 80 & 443, which was the solution I guess. I dont know how I could have overlooked that for 2 days, but it is what it is (sorry for my stupidity).

I still have no idea why the apps could not reach each other before, but maybe the system restart fixed that.

I would also still like to know if this approach is right or if there is a better way. Also I normally try to only use DHCP for static IPs via the unifi client device settings, so that the IP assignment is centralized. With this solution that seems not possible since only one interface can have DHCP and for me that is br0 where I have a VM on.

1 Like