Hi,
Let’s assume that there is a 2Master+3Worker node based cluster is setup in a lab in India and some k8s services/pods are running and providing service to the customer.
Now, if I want to shutdown the whole cluster (all nodes including master and worker) and ship it to a place in US or Sweden and then start the whole cluster again.
The site in US/Sweden will have a different IP/network than in India. Does this change in IP network has any impact on the k8s cluster. Is it possible to prepare the cluster in the premises of a development center in India, shut them down and then start them up at different deployment locations with different time/zone/IP.
Thanks
Deb
This will be a bit tricky. Most of the changes you’ll need to make will be in the kube-apiserver.yaml file, although since your cluster has two masters, you will need to configure the load balancer you have in front of the apiserver instances. You might need to reinstall the CNI plugin as well. Definitely back the server up before doing this, especially etcd.
Thanks for giving the lead @rob_kodekloud
Do you by any chance has any link/material for the steps to achieve this.
Most of this you’ll find in the reference for kube-apiserver. A lot of this depends upon how you’re going to wire your cluster up and connect it to the new network. You have:
- The IP of the load balancer (kube-apiserver)
- The IPs of the ported nodes, including the masters (kube-apiserver for the latter)
- The service CIDR (which hopefully you won’t have to change; kube-apiserver)
- The CIDR of the pod network (hopefully you won’t have to change this either; depends on the CNI you’re using).
There isn’t going to be guide for this, since your use-case is a little unusual. You’re going to need to be comfortable updating the networking settings for the ported nodes, trying to network them between themselves w/o reference to the new host network. Might take you a while to get everything working again.