Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-34110

[Backport 4.12] The ingress-operator spuriously updates ingressClass on startup

XMLWordPrintable

    • Low
    • No
    • 1
    • Sprint 254
    • 1
    • Rejected
    • False
    • Hide

      None

      Show
      None
    • Hide
      * Previously, the Ingress Operator did not specify `ingressClass.spec.parameters.scope`, while the Ingress Class API object specifies type `cluster` by default. This caused unnecessary updates to all Ingress Classes when the Operator starts. With this update, the Ingress Operator specifies `ingressClass.spec.parameters.scope` of type `cluster`. (link:https://issues.redhat.com/browse/OCPBUGS-34110[*OCPBUGS-34110*]) Backport to 4.12.

      Cause: The ingress-operator does not specify ingressClass.spec.parameters.scope, while the Ingress Class API object specifies "Cluster" by default"
      Consequence: Unneeded updates to all Ingress Classes when the ingress-operator starts.
      Fix: Ingress-operator now specifies ingressClass.spec.parameters.scope of type "Cluster".
      Result: No more unneeded updates of Ingress Classes.
      Show
      * Previously, the Ingress Operator did not specify `ingressClass.spec.parameters.scope`, while the Ingress Class API object specifies type `cluster` by default. This caused unnecessary updates to all Ingress Classes when the Operator starts. With this update, the Ingress Operator specifies `ingressClass.spec.parameters.scope` of type `cluster`. (link: https://issues.redhat.com/browse/OCPBUGS-34110 [* OCPBUGS-34110 *]) Backport to 4.12. Cause: The ingress-operator does not specify ingressClass.spec.parameters.scope, while the Ingress Class API object specifies "Cluster" by default" Consequence: Unneeded updates to all Ingress Classes when the ingress-operator starts. Fix: Ingress-operator now specifies ingressClass.spec.parameters.scope of type "Cluster". Result: No more unneeded updates of Ingress Classes.
    • Bug Fix
    • In Progress

      This is a clone of issue OCPBUGS-6701. The following is the description of the original issue:

      Description of problem:

      The function desiredIngressClass doesn't specify ingressClass.spec.parameters.scope, while the ingressClass API object specifies "Cluster" by default.
      
      This causes unneeded updates to all IngressClasses when the CIO starts. The CIO will fight with the API default any time an update triggers a change in an IngressClass.

      Reference: https://github.com/kubernetes/api/blob/master/networking/v1/types.go#L640 

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

      4.8+

      How reproducible:

       

      Steps to Reproduce:

      We really need https://issues.redhat.com/browse/OCPBUGS-6700 to be fixed before we can identify these spirituous updates. But when it is fixed:

      # Delete CIO
      oc delete pod -n openshift-ingress-operator  $(oc get -n openshift-ingress-operator pods --no-headers | head -1 | awk '{print $1}')
      
      # Wait a minute for it to start back up
      # Should be NO updates to IngressClasses
      oc logs -n openshift-ingress-operator $(oc get -n openshift-ingress-operator pods --no-headers | head -1 | awk '{print $1}') -c ingress-operator | grep "updated IngressClass"
      
      # Instead, we see this every time CIO starts up
      2023-01-26T20:57:15.281Z    INFO    operator.ingressclass_controller    ingressclass/ingressclass.go:63    updated IngressClass    {"name": "openshift-default",  

      Actual results:

      2023-01-26T20:57:15.281Z    INFO    operator.ingressclass_controller    ingressclass/ingressclass.go:63    updated IngressClass    {"name": "openshift-default", ...

      Expected results:

      No update to ingress upon CIO restart

      Additional info:

       

              cholman@redhat.com Candace Holman
              openshift-crt-jira-prow OpenShift Prow Bot
              Hongan Li Hongan Li
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

                Created:
                Updated:
                Resolved: