Keep existing tailscale subnet routes when deploying and deleting

[manuelbuil]

Is your feature request related to a problem? Please describe.

When deploying k3s with tailscale, we are resetting the tailscale subnet routes and adding the ones required for flannel communication. Therefore, if a user is already operating tailscale with subnet routes, his/her traffic will get disrupted.

When removing, we are deleting all subnet-routes, so we should also find a way to fix this

Describe the solution you'd like

Before resetting, we should read the current subnet routes, save them, and then, after the reset add them back together with the new one for flannel

Describe alternatives you've considered

Additional context

[DCjanus]

It seems that Tailscale does not appear to expose the interface of the current configuration. Is there any good way to solve this problem?

[manuelbuil]

It seems that Tailscale does not appear to expose the interface of the current configuration. Is there any good way to solve this problem?

What do you mean by the interface of the current configuration? AFAIK, there is only one tailscale interface and if you change the config, that interface will use the new configuration

[DCjanus]

Sorry for my mistake, I mean there is no API from tailscale to get current subroute config. And tailscale set would overwrite current subroutes.

I also encountered this issue. I tried to fix it, but I couldn't find a way to get the current subnet router.

[manuelbuil]

Sorry for my mistake, I mean there is no API from tailscale to get current subroute config. And tailscale set would overwrite current subroutes.

I also encountered this issue. I tried to fix it, but I couldn't find a way to get the current subnet router.

I thought this output would return the subroute config https://github.com/k3s-io/k3s/blob/master/pkg/vpn/vpn.go#L123

[DCjanus]

You are right, This is really good news.