Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-24044

Converting load balancer service from internal scope to external keeps internal load balancer IP on GCP

XMLWordPrintable

    • Critical
    • No
    • CLOUD Sprint 246
    • 1
    • Proposed
    • False
    • Hide

      None

      Show
      None
    • Release Note Not Required
    • In Progress

      Reproducer:
      1. On a GCP cluster, create an ingress controller with internal load balancer scope, like this:

      apiVersion: operator.openshift.io/v1
      kind: IngressController
      metadata:
        name: foo
        namespace: openshift-ingress-operator
      spec:
        domain: foo.<cluster-domain>
        endpointPublishingStrategy:
          type: LoadBalancerService
          loadBalancer:
            dnsManagementPolicy: Managed
            scope: Internal
      

      2. Wait for load balancer service to complete rollout

      $ oc -n openshift-ingress get service router-foo
      NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
      router-foo LoadBalancer 172.30.101.233 10.0.128.5 80:32019/TCP,443:32729/TCP 81s
      

      3. Edit ingress controller to set spec.endpointPublishingStrategy.loadBalancer.scope to External

      the load balancer service (router-foo in this case) should get an external IP address, but currently it keeps the 10.x.x.x address that was already assigned.

            joelspeed Joel Speed
            rfredett@redhat.com Ryan Fredette
            Zhaohua Sun Zhaohua Sun
            Votes:
            0 Vote for this issue
            Watchers:
            10 Start watching this issue

              Created:
              Updated:
              Resolved: