0

Done a lot of googling and reading, as well as viewing as a guest from Google searches on this topic as my goal is to setup a small kube lab cluster from scratch myself. Decided to sign up today and ask for assistance from those who've perhaps done this process recently.

Managed to get cri-o and kube both v1.20 installed on my master node, and so just created 1 worker node so far for proof of concept as I'm on trial plan so don't want to fire up unnecessary extra VMs/resources until I'm sure I've got the whole process working.

At this point cri-o status is ok, but kubelet on master VM is crashing with exit code 255 probably because I haven't run kubeadm init yet.

My question now is: I've reserved ephemeral internal ip's for the VMs but should I reserve external IP's for master? Yes ssh will be easier, but any limitation on kube running if I don't?

Thanks and any further tips about creating kubeadm init cluster parameters for my cloud env shall be gratefully received.

p.s my VMs are E2 with 2 vcpu's, 4GB RAM, Ubuntu 20.04 minimal

Kub-cluster
  • 1
  • Kelsey Hightower published "Kubernetes the hard way", great instructions how to install Kubernetes from scratch. – berndbausch Mar 21 '21 at 01:31
  • Answering your question about internal/external IP's allocation: This would be related to the network configuration that you are having and how your cloud provider is exposing your VM's outside (access to ports on your `VM`). Saying to create a cluster from scratch is most likely using [Kubernetes: The hard way](https://github.com/kelseyhightower/kubernetes-the-hard-way) and there is no `kubeadm` part in it. If you need tips on `kubeadm` the official documentation is a good place to start. Am I correct in saying that you are trying to run a cluster with `kubeadm` on EC2 (AWS)? – Dawid Kruk Mar 22 '21 at 15:04

0 Answers0