Can't handle loadBalancerIP change

[ouyangde]

Describe the bug When change loadBalancerIP of a service, VIP not get reassigned.

To Reproduce Steps to reproduce the behavior:

1. create a deployment of nginx.
2. create service/http1 for that nginx, with type: LoadBalancer and loadBalancerIP: 192.168.0.128
3. curl http://192.168.0.128/ return normal contents.
4. change loadBalancerIP of service/http1 to 192.168.0.126
5. curl http://192.168.0.126/ Connection refused
6. delete service/http1 and recreate it with type: LoadBalancer and loadBalancerIP: 192.168.0.126
7. curl http://192.168.0.126/ return normal contents.

Expected behavior step 5 should return normal contents.

Environment (please complete the following information):

• OS/Distro: [e.g. Centoos 7.8]
• Kubernetes Version: [e.g. v.1.23]
• Kube-vip Version: [e.g. 0.5.12]

Kube-vip.yaml: spec: containers:

• args:
  • manager env:
  • name: vip_arp value: "true"
  • name: port value: "6443"
  • name: vip_interface value: eth1
  • name: vip_cidr value: "32"
  • name : lb_enable value: "true"
  • name : lb_port value: "6443"
  • name: lb_fwdmethod value: local
  • name: cp_enable value: "true"
  • name: cp_namespace value: kube-system
  • name: vip_ddns value: "false"
  • name: svc_enable value: "true"
  • name: svc_election value: "true"
  • name: vip_leaderelection value: "true"
  • name: vip_leaseduration value: "5"
  • name: vip_renewdeadline value: "3"
  • name: vip_retryperiod value: "1"
  • name: address value: 192.168.0.100
  • name: prometheus_server value: :2112

Additional context kube-vip deploy method: static pods number of kube-vip instance: 3

[todaygood]

I need this feature too.

[lubronzhan]

It's already supported. But for new kube-vip you need to modify the Annotation kube-vip.io/loadbalancerIPs now, not the spec.loadBalancer. Since the spec.loadBalancer was deprecated in k8s 1.24

metadata:
  annotations:
    kube-vip.io/loadbalancerIPs: 10.92.96.194