Instruction are a bit outdated

vano_mashey

In the Lab of chapter 3 the instructions to work with calico network provider are outdated. Old version of calico is used 2.6.
On the calico site the suggest version is 3.3. Standard instruction do not work as etcd and rbac have to be deployed separately.
Using instuctions in the Lab does not give much success. coredns and calico pods are failing. Migrating to 3.3 version brings issues with ipvs, which is aparently new thing.

Please advice if there is any instuctions/updates on the current version.

vano_mashey

It could be because the default version installed in centos is 1.12.2 and not 1.11

chrispokorni

Hi, I believe this is the reason why the lab gives specific instructions about the kubernetes (kubeadm, kubelet, kubectl) and calico versions to install, to avoid headaches. I just completed lab 3 with k8s 1.11 and calico 2.6, and it all worked fine.
I am running my cluster on GCP VMs at 2 vCPUs with Ubuntu 16. I also created a firewall rule to enable all traffic (all protocols, all ports, all sources) because by default the high ports used by k8s may be blocked.
Regards,
-Chris