Cobia - network setup bug report

[nikkon]

Hi all,

I run TrueNAS-SCALE-23.10.2. As I have 2 NICs I run a bond0 LACP setup. It works great but I cannot force my containers to get LAN DHCP IPs.
So I decided to create a bridge interface (br0) and add the bond to it. the bridge has no IP assigned to it.
I needed to restart to make the setup works. After the restart I lost access to the my nas. I fixed it via IMPI but I think there is a bug here.
I also tried to delete the bridge - and this is not an option that works in the cli menu. Bug number 2.
Any thoughts?

 

[Fleshmauler]

I could be wrong, but wouldn't you want to do this in the following steps to get it working?

1. create the bridge
2. Set the bond as a member of the bridge
3. remove ips from the bond
4. insert removed ips to the bridge

 

[nikkon]

That explains my mistake. I kept the IP on the bind not the bridge

 

[nikkon]

so - part of the problem solved.
I still cannot make the containers get a Lan Ip.
I modified the K8s settings acording to my Lan specs :

 

[Fleshmauler]

Not sure why you'd want your container to have a LAN IP, if you could explain use case or specifically what apps you're using maybe I could be of further use. When I used to bother messing around with Node IP etc I would have to unset & reset the pool to make it take effect - maybe you're missing that step?

Using only Plex & unifi I'm able to use both without any issues by using Truenas IP + relevant port.

Here are my settings for example:

after that for example https://LAN IP GOES HERE:32400 takes me to my plex interface (not sure why I'm bothering to censor my IP on a LAN subnet as if it matters).

 

[nikkon]

I need my Plex and photo prism to have a lan IP.

 

[Fleshmauler]

Hear me out; wouldn't Plex just use your Truenas LAN ip + whatever port you assigned to it & the cluster IP is completely irrelevant since it is hosted locally on the Truenas device & Truenas would immediately understand the request for whatever port is assigned to the app & do the needful? I'm 100% going to assume that photo prism will work the same way.

For example lets assume Truenas is setup on LAN IP 192.168.0.2 & you have plex open on default Plex port 32400... wouldn't you just use 192.168.0.2:32400 in a browser to access Plex? I've never had issues Plex being detected on my LAN (browser, app, or HTPC) like this or remotely even though my cluster IP is whatever 172.default-nonsense listed above.

Please let me know if I'm misunderstanding.

 

[nikkon]

I am not saying is not working like this. it does today. what I'm saying is - I want to separate them by Ip and I can't

 

[Fleshmauler]

Next question, and someone smarter than me could correct me; wouldn't this just setup the cluster IP on LAN & instead of using the Truenas IP + Set Port to connect to the apps, you'd use Whatever Cluster LAN IP + Port to connect to the apps? If so, then neitherof them would ever truly get 'their own IP'.

At best all this would accomplish is separate the LAN IP of the cluster from your Truenas Scale LAN IP. I guess you could have a cluster per APP with their own IPs & that would halfway accomplish what you a trying to do? In that Case we could have Truenas on 192.168.0.2, Plex Cluster on .0.3, Prism Cluster on .0.4, and you then connect to Plex via 192.168.0.2:32400 & Prism on 192.168.0.3:whatever for example?

I'm still not 100% on why we'd want this, but hopefully this points you to your end goal.