This contains some example code to determine how long "failovers" are taking within kube-vip, the server component should live within the cluster and the client should be externally.
Simply apply the manifest to a working cluster that has kube-vip deployed:
kubectl apply -f ./demo/server/deploy.yaml
Retrieve the loadBalancer IP that is fronting the service:
kubectl get svc demo-service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
demo-service LoadBalancer 10.104.18.147 192.168.0.217 10002:32529/UDP 117m
From elsewhere, clone the kube-vip repository and connect the client to the server endpoint (loadBalancer IP) with the following command:
go run ./demo/client/main.go -address=<vip>
You will only see output when the client has reconcilled the connection to a pod beneath the service, where it will print the timestamp to reconnection along with the time in milliseconds it took:
15:58:35.916952 3008
15:58:45.947506 2005
15:58:57.983151 3007
15:59:08.013450 2005
15:59:20.046491 3008
15:59:30.076341 2507
15:59:42.110747 3008
On a machine or control plane that has kubectl and has the credentials to speak to the cluster we will run a command to find the demo pod and kill it every 10 seconds:
while true ; do kubectl delete pod $(kubectl get pods | grep -v NAME | grep vip| awk '{ print $1 }'); sleep 10; done