Lab 2.2. Deploy a New Cluster

stanislawfreund

I am trying the Lab Exercise, but after executing the command:
'bash k8scp.sh | tee $HOME/cp.out' at some point I am asked to restart the system. The command is already finished prior to the restart, but I never see the output as pictured in the lab excersice:

So I don´t even know if the installation is done correctly.

Can anybody help me?

Kind regards!

chrispokorni

Hi @stanislawfreund,

What user executes the command? What guest OS distribution is running your VM? Is it a local or cloud VM?
Please provide any errors or warning messages generated while the script is running.

Regards,
-Chris

stanislawfreund

Hi @chrispokorni,
I am using a two Ubuntu VM on Virtualbox. One for the control plane and the other one for the worker node. The guest os on both is Ubuntu 20.04. The commands were executed without sudo, by the regular user. For now I could execute both k8s.sh on the cp as well as k8sWorker.sh on the Worker node. But I can´t join the nodes.

Do I have to execute the command kubectl apply -f podnetwork.yaml? I Saw that we already installed Cilium with k8s.sh.

Kind regards,
Stanislaw

chrispokorni

Hi @stanislawfreund,

The connection refused errors could be triggered if the promiscuous mode is not enabled to allow all traffic in VirtualBox for each VM. A single bridged network interface is recommended per VM.

Running the sudo kubeadm join ... repeatedly on the same host does not help. Instead it causes many more errors and failures. Please run a sudo kubeadm reset on the worker node prior to re-running the join command.

If followed closely, the lab guide steps should be sufficient to bootstrap a working Kubernetes two-node cluster. There is no need to re-run steps or to run them out of order to fix accumulated errors.

Regards,
-Chris