This message was deleted.
# rke2
a
This message was deleted.
c
did you forget to set the CNI to cilium on those nodes? honestly I wouldn’t have expected them to even join the cluster if they don’t have the same CNI selected.
It kinda sounds like they’re not even in the same cluster
b
No all nodes have the same config.yaml dropped on them
c
or are you using rancher to do this?
b
No this is the rancher cluster we're bootstrapping
c
did you drop the config after starting rke2? It really sounds like you built 3 different single-node clusters.
b
Nope before... At least per how this is laid out
So kubectl get nodes shows all three node
c
huh. Did you bring it up with the default CNI and then switch to cilium without cleaning up in between?
b
It's just cilium is having a fit because there is a flannel.1 iface
No we're bootstrapping with cni: cilium
So like the first time we start rke2 the config.yaml already exists in /etc
c
well something created the flannel interface, and I can’t think of what that would be other than flannel having been running on these nodes at some point.
b
Yea it's a rocky node so the only way I can think it would be getting there is via canal... It's probably the automation. But it's not obvious where
c
I guess I’d probably just step through the tf and ensure that things are actually happening in the order you think they should be happening
b
Think we got it... Totally the terraform. Thanks for the help Brad!
👍 1