Is there any documentation (e.g. sequence diagrams?) available on the “node join” process for a custom / RKE cluster? I’ve looked through the Rancher docs, but didn’t find anything more detailed than how to copy/paste the command.
I’m curious as to what launches kubelet / etcd / etc. I’ve noticed the share-mnt container polls very early on waiting for kubelet to start, but is it launched by rancher-server making a “back-door” connection to a privileged container? Is there another local container that manages the hyper-kube K8s containers?
Any insight would be helpful! Thanks!