Upgrading PKS with NSX-T from 1.0.x to 1.1
Yesterday, Pivotal Container Service 1.1 dropped and, as it’s something I’ve been actively learning in my lab, I wanted to jump on the upgrade straight away. PKS with NSX-T is a really hot topic right now and I think it’s going to be a big part of the future CNA landscape.
My Lab PKS 1.0.4 deployment is configured as a “NO-NAT with Logical Switch (NSX-T) Topology” as depicted in the diagram below (from the PKS documentation). My setup has these network characteristics:
- PKS control plane (Ops Manager, BOSH Director, and PKS VM) components are using routable IP addresses.
- Kubernetes cluster master and worker nodes are using routable IP addresses.
- The PKS control plane is deployed inside of the NSX-T network. Both the PKS control plane components (VMs) and the Kubernetes Nodes use routable IP addresses.
I used William Lam’s excellent series on PKS with NSX-T to configure a lot of the settings, so I am going to assume a familiarity with that series. If not, I suggest you start there to get a good understanding of how everything is laid out.