A demo on PodDisruptionBudget

This is a very simple project to demonstrate the capability of PodDisruptionBudget.

Prerequisite

You will need to prepare a cluster beforehand. The cluster can be an actual k8s cluster running in the cloud, or a minikube/k3s/kind cluster. The cluster should at least have 2 nodes.

Usage

1. Apply manifests/deployment.yaml
2. You should have two nginx pods running.
3. Use kubectl get pod -o wide to verify if they are scheduled on the same node.
   1. They should be.
4. Wait until both of them are 1/1.
5. Open up another terminal, and run watch kubectl get pod.
   1. You should be able to see a list of all running pods.
   2. The list should be refreshing once every few seconds.
6. Now, take note of the node (node 1 from now on) on which those two pods are scheduled.
7. Do kubectl drain {{_node 1_}}.
   1. You may need to supply --ignore-daemonsets --force
8. Take notice of the result of watch kubectl get pod in another terminal.
   1. Both pods should restart and become 0/1
   2. If you try to reach the pods either through a service or by port-forwarding, you should receive Connection reset.
   3. After a while, those two pods should be respawned on the other node (node 2)
9. Now, recover the state of node by kubectl undrain {{_node 1_}}
10. kubectl apply -f manifests/pdb.yaml
11. Do kubectl drain {{_node 2_}}.
12. You should see the effect of PodDisruptionBudget now.