Multiple network default routes in Scale?

I recently installed Scale after years of running Core - noticed this interesting situation.

Looking further at the UI, I can’t see how one would specify two default routes, leading me to suspect this might be a learned route?

Lastly, for completness, sharing my network interface config.

How do two default routes happen? I remember jgreco explaining on the old forums why you can only have one default gateway and that explanation made sense to me. Now I see this and I am curious what is the semantics of this (and how do the Linux/BSD network stacks differ in this regard).

1 Like

One for each network as you have multiple NICs?

What does the output of “ip route” show?

One for each network as you have multiple NICs?

I’ll see if I can dig up the original article, but I thought you could only have 1 default gateway specified (at least on FreeBSD)?

What does the output of “ip route” show?

Here is the ip route output - there seems to be a single default specified.

admin@storage[~]$ ip route
default via dev enp134s0f0np0 proto static dev enp7s0f1 proto kernel scope link src dev enp134s0f0np0 proto kernel scope link src dev enp134s0f1np1 proto kernel scope link src dev eno1 proto kernel scope link src dev kube-bridge proto kernel scope link src

Hm, perhaps setting this would be the culprit?

EDIT: the reason why I thought of specifying those values in Kubernetes settings is here.

Well, I know you can’t have 2 actual default routes, but, wasn’t sure what Scale was reporting since it showed two. ip route shows you have 1 which is correct, 10,0.50.2 is the one default, so, that matches your networking setup. Therefore, it has to be because of your so called kubernetes settings, is that what you actually wanted? It shouldn’t be calling it a default route though.

Yeah, if you follow the link in the above post, you’ll find more details on why I did that. In short, I had hoped that by providing these two, the K8s traffic will be routed through a dedicated interface and subnet. That didn’t happen, however - still waiting on some thoughts as to why.

Still, I think that additional value in the Default Route is very confusing and should not be there.

I don’t think it should be there either, if they want it there, they should change the name from default route as that has a specific meaning. I think it’s a bug.

I didn’t read the articles. But I don’t know specifically how to route your Kubernetes traffic through the dedicated interface so someone else needs to chime in. There may be some specific rules there.

1 Like

I’ll file this as a bug. Thank you for your help!