Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-59176

[OCPFeatureGate: GCPClusterHostedDNSInstall] gcp-custom-dns Job Permafailing Several DNS Tests

XMLWordPrintable

    • Quality / Stability / Reliability
    • False
    • Hide

      None

      Show
      None
    • 2
    • Moderate
    • No
    • None
    • Rejected
    • NI&D Sprint 275, NI&D Sprint 276, NI&D Sprint 277, NI&D Sprint 278
    • 4
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Description of problem:

          Some e2e tests are failing with the job "gcp-custom-dns" (was "gcp-user-provisioned-dns") for featuregate "GCPClusterHostedDNSInstall" which is promoted to GA in 4.20.
      
      For more details about "GCPClusterHostedDNSInstall" please refer to https://github.com/openshift/enhancements/pull/1468/files#diff-230cf161d0e8b0dc075e6760516b1171448adae3e6ee937ac9a97a0ce44862ddR106-R120
      
      In short, the http2 and grpc tests use dedicated ingresscontrollers, and gateway also has separated LB and dnsrecord, so the default wildcard created by the new static CoreDNS won't work for those tests.
      
      
      And another issue is: grpc/http2 tests use ingressDomain to generate shardFQDN so the shard ingresscontroller domain looks like:
      
       "domain: e2e-test-router-http2-vtzjw.apps.hongli-gcp20.qe.gcp.devcluster.openshift.com"
      
      If the dnsrecords of shard cannot be published, the "*.e2e-test-router-http2-vtzjw.apps.<baseDomain>" is resolved to "*.apps.<baseDomain>", so the request goes to default router and causes unexpected result.
      
      
      Here are the failing tests:
      [sig-network-edge][Conformance][Area:Networking][Feature:Router][apigroup:route.openshift.io][apigroup:config.openshift.io]
       The HAProxy router should pass the http2 tests 
      [apigroup:image.openshift.io][apigroup:operator.openshift.io] 
      [Suite:openshift/conformance/parallel/minimal] 
                  
      [sig-network-edge][Conformance][Area:Networking][Feature:Router] The 
      HAProxy router should pass the gRPC interoperability tests 
      [apigroup:route.openshift.io][apigroup:operator.openshift.io] 
      [Suite:openshift/conformance/parallel/minimal] 
      
      [sig-network-edge][OCPFeatureGate:GatewayAPIController][Feature:Router][apigroup:gateway.networking.k8s.io]
       Ensure LB, service, and dnsRecord are created for a Gateway object 
      [Suite:openshift/conformance/parallel]             
                    : 
      [sig-network-edge][OCPFeatureGate:GatewayAPIController][Feature:Router][apigroup:gateway.networking.k8s.io]
       Ensure HTTPRoute object is created 
      [Suite:openshift/conformance/parallel] 
      
      

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

          4.20

      How reproducible:

          100%

      Steps to Reproduce:

          1. run e2e test with job "e2e-gcp-custom-dns" (was "e2e-gcp-user-provisioned-dns")
          2. 
          3. 
          

      Actual results:

          tests fail

      Expected results:

          1. the sig-network-edge tests should pass on "gcp-custom-dns" job
          2. shardFQDN should use oc.Namespace() + "." + BaseDomain

      Additional info:

          example job link: https://prow.ci.openshift.org/view/gs/test-platform-results/pr-logs/pull/openshift_installer/9792/pull-ci-openshift-installer-main-e2e-gcp-custom-dns/1957526619325206528 
      
          code of shardFQDN: https://github.com/openshift/origin/blob/main/test/extended/router/http2.go#L257

              rhn-support-hongli Hongan Li
              rhn-support-hongli Hongan Li
              None
              None
              Hongan Li Hongan Li
              None
              Votes:
              0 Vote for this issue
              Watchers:
              10 Start watching this issue

                Created:
                Updated: