Home / server / Questions / 1088836
Accepted
spoovy
Asked: 2022-01-04 01:51:33 +0800 CST

K8s - How trigger refresh of LoadBalancer svc IP?

  • 772

I'm using metallb to assign IPs to LoadBalancers in a home lab k8s cluster. I've had to shrink the IP pool though as I need to use some of them in another test cluster.

I've done this but the IPs of the existing LB svcs have remained the same afterwards -- I need to give the LB svcs a kick somehow to get them to request new IPs. How can I do this?

kubernetes

1 Answers

  1. Best Answer

    It looks like the answer to your question is contained in this github thread.

    Everything should come down to recreate your service with the appropriate configmap. Look at this answer:

    Before change

    [root@m-k8s ~]# kubectl get services
NAME         TYPE           CLUSTER-IP      EXTERNAL-IP    PORT(S)        AGE
cfgmap-svc   LoadBalancer   10.100.199.95   192.168.1.11   80:31463/TCP   2m5s
kubernetes   ClusterIP      10.96.0.1       <none>         443/TCP        4h13m

    Change IP address

    [root@m-k8s ~]# cat metallb-l2.yaml | grep 192.168
      - 192.168.1.11-192.168.1.13
[root@m-k8s ~]# sed -i 's/11/101/;s/13/103/' metallb-l2.yaml
[root@m-k8s ~]# cat metallb-l2.yaml | grep 192.168
      - 192.168.1.101-192.168.1.103

    Just apply
    kubectl apply -f metallb-l2.yaml

    Delete and revise by kubelet

    [root@m-k8s ~]# kubectl delete pods --all -n metallb-system
pod "controller-65895b47d4-l5xnn" deleted
pod "speaker-bk8gr" deleted
pod "speaker-bksns" deleted
pod "speaker-c2z6z" deleted
pod " speaker-chlt5" deleted
pod "speaker-scx6x" deleted

    Recreate svc

    [root@m-k8s ~]# kubectl delete service cfgmap-svc
service "cfgmap-svc" deleted
[root@m-k8s ~]# kubectl expose deployment cfgmap --type=LoadBalancer --name=cfgmap-svc --port=80
service/cfgmap-svc exposed

    Check svc

    [root@m-k8s ~]# kubectl get svc
NAME          TYPE           CLUSTER-IP       EXTERNAL-IP     PORT(S)        AGE
cfgmap-svc3   LoadBalancer   10.104.192.184   192.168.1.101   80:31440/TCP   8s
kubernetes    ClusterIP      10.96.0.1        <none>          443/TCP        4h34m

    You can also look at this github issue. Note, it is old thread.

    • 1