Get cluster started ...

[liszca]

Following the documentation I setup my cluster like this:

pi5n1 - 192.168.1.31
pi5n2 - 192.168.1.32
pi5n3 - 192.168.1.33

curl -sfL https://get.k3s.io | sh -
for i in 192.168.1.3{2..3};do ssh $i "curl -sfL https://get.k3s.io | K3S_URL=https://192.168.1.31:6443 K3S_TOKEN=NotShownHere::server:NotShownHere sh -";done

This results in kubectl get nodes:

How should kubectl get nodes look with 3 nodes High-Availability?
How can I have multiple control-planes (on all three Pis as fallback)?
How to setup etcd?
Can't run kubectl edit svc kubernetes on any pi but pi5n1

[brandond]

Please read the docs: https://docs.k3s.io/datastore/ha-embedded

Only server nodes have the admin kubeconfig. Agent nodes are less privileged and do not have a copy. See also https://docs.k3s.io/cluster-access

[liszca]

Is it actually necessary to provide a node-token on --cluster-init?

[brandond]

This is covered in the docs...
https://docs.k3s.io/cli/token#server

If no token is provided when starting the first server in the cluster, one is created with a random password.

Please go read the docs.