Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-48115

The apiservice v1.admission.autoscaling.openshift.io gets not reachable after deleting the ClusterResourceOverride for clusterresourceoverride-operator

XMLWordPrintable

    • Quality / Stability / Reliability
    • False
    • Hide

      None

      Show
      None
    • 3
    • Important
    • None
    • None
    • None
    • AUTOSCALE - Sprint 267, AUTOSCALE - Sprint 270
    • 2
    • Done
    • Bug Fix
    • Hide
      * Previously, if you deleted the `clusterresourceoverride-operator` service or uninstalled the Cluster Resource Admission Override Operator, the `v1.admission.autoscaling.openshift.io` API service was unreachable and prevented needed cluster functionas, such as installing other Operators on the cluster. With this release, a fix ensures that if the Cluster Resource Admission Override Operator is uninstalled, the `v1.admission.autoscaling.openshift.io` API service is also deleted so that the cluster functions are not impacted. (link:https://issues.redhat.com/browse/OCPBUGS-48115[OCPBUGS-48115])
      Show
      * Previously, if you deleted the `clusterresourceoverride-operator` service or uninstalled the Cluster Resource Admission Override Operator, the `v1.admission.autoscaling.openshift.io` API service was unreachable and prevented needed cluster functionas, such as installing other Operators on the cluster. With this release, a fix ensures that if the Cluster Resource Admission Override Operator is uninstalled, the `v1.admission.autoscaling.openshift.io` API service is also deleted so that the cluster functions are not impacted. (link: https://issues.redhat.com/browse/OCPBUGS-48115 [ OCPBUGS-48115 ])
    • None
    • None
    • None
    • None

      Description of problem:

      After the clusterresourceoverride-operator executed test case OCP-27070(install clusterresourceoverride-operator, create a ClusterResourceOverride and then delete it), the apiservice v1.admission.autoscaling.openshift.io will become not reachable. As a result, some other operators will not be able to install successfully due to "unable to retrieve the complete list of server APIs:[ admission.autoscaling.openshift.io/v1|http://admission.autoscaling.openshift.io/v1]: stale GroupVersion discovery:"  error.

      Version-Release number of selected component (if applicable):

      4.17.11 + ClusterResourceOverride Operator.v4.17.0-202412170235    

      How reproducible:

      Always    

      Steps to Reproduce:

      1. Install ClusterResourceOverride Operator.v4.17.0-202412170235  
      2. Create ClusterResourceOverride with below yaml file:
      apiVersion: operator.autoscaling.openshift.io/v1
      kind: ClusterResourceOverride
      metadata:
          name: cluster
          namespace: clusterresourceoverride-operator 
      spec:
        podResourceOverride:
          spec:
            memoryRequestToLimitPercent: 50 
            cpuRequestToLimitPercent: 25 
            limitCPUToMemoryPercent: 200 
      
      3. Delete ClusterResourceOverride:
      % oc delete ClusterResourceOverride cluster -n clusterresourceoverride-operator  
      

      Actual results:

      After ClusterResourceOverride deleted, the service clusterresourceoverride got deleted, the apiservice v1.admission.autoscaling.openshift.io become not reachable. As a result, some other operators will not be able to install successfully due to "unable to retrieve the complete list of server APIs:[ admission.autoscaling.openshift.io/v1|http://admission.autoscaling.openshift.io/v1]: stale GroupVersion discovery:"  error.

       

      % oc describe apiservice v1.admission.autoscaling.openshift.io
      Name:         v1.admission.autoscaling.openshift.io
      Namespace:    
      Labels:       operator.autoscaling.openshift.io/clusterresourceoverride=true
      Annotations:  service.alpha.openshift.io/inject-cabundle: true
      API Version:  apiregistration.k8s.io/v1
      Kind:         APIService
      Metadata:
        Creation Timestamp:  2025-01-08T02:49:59Z
        Resource Version:    48448
        UID:                 d9b3485c-d4d2-485e-821a-6ed669de0642
      Spec:
        Ca Bundle:               xxxx
        Group:                   admission.autoscaling.openshift.io
        Group Priority Minimum:  1000
        Service:
          Name:            clusterresourceoverride
          Namespace:       clusterresourceoverride-operator
          Port:            443
        Version:           v1
        Version Priority:  15
      Status:
        Conditions:
          Last Transition Time:  2025-01-08T02:50:02Z
          Message:               service/clusterresourceoverride in "clusterresourceoverride-operator" is not present
          Reason:                ServiceNotFound
          Status:                False
          Type:                  Available
      Events:                    <none>
      % oc get service
      No resources found in clusterresourceoverride-operator namespace. 
      

      Expected results:

      If the operator removes the service, it should also remove the API service. So other operations won't be impacted    

      Additional info:

      logs is available at https://drive.google.com/drive/folders/1PqpvrOhRpn97Z5Wf1_8G0CpZrvoxWfKq    

              rh-ee-macao Max Cao
              xiyuan@redhat.com Xiaojie Yuan
              None
              None
              Paul Rozehnal Paul Rozehnal
              None
              Votes:
              0 Vote for this issue
              Watchers:
              7 Start watching this issue

                Created:
                Updated:
                Resolved: