I have k3os version v0.21.5-k3s2r1
k3os installed Traefik version v1.7.19 by default
DaemonSet of Traefik created svclb-traefik pods for all nodes and remain in pending state
I typically see this when some other process is already listening on the entrypoint ports for traefik (these are usually 80 and 443).
Did you look at why those pods are failing to start?
kubectl -n kube-system describe pod/one-of-those-pod-names
It should give you an indication if this is the problem: tracking down and stopping the offending process can sometimes be a little more difficult depending on whether it's something outside of k8s (apache was already using those ports), or something inside of k8s (a daemonset of svclb pods in -another- namespace has already started up and is listening on those ports). I typically have to delete the pods in the offending namespace, and then traefik will sneak in and take over those ports before the others can restart. This leaves the svclb that I -don't- want in a Pending state instead.