Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-18378

ovn routingViaHost-true flag breaks connectivity to pod services from hostnetwork pods

XMLWordPrintable

    • +
    • Critical
    • Yes
    • SDN Sprint 241
    • 1
    • Approved
    • False
    • Hide

      LGW installation fails on OVNK.

      Show
      LGW installation fails on OVNK.
    • Hide
      9/6 downstream fix merged. Green
      9/5: upstream fix merged, downstream merge PR open. Green
      Show
      9/6 downstream fix merged. Green 9/5: upstream fix merged, downstream merge PR open. Green

      Description of problem:

      OVN local mode breaks connectivity to pod services from hostnetwork pods. As a result metallb/sriov and several other operators don't work.

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

      4.14.0-0.nightly-2023-08-28-154013

      How reproducible:

      Set routingViaHost: true flag in following resource networks.operator.openshift.io cluster
      
      Try to apply sriov or metallb policy and hit following error:
      
      Internal error occurred: failed calling webhook "operator-webhook.sriovnetwork.openshift.io": failed to call webhook: Post "https://operator-webhook-service.openshift-sriov-network-operator.svc:443/validating-custom-resource?timeout=10s": context deadline exceeded
      
      OR
      
       Internal error occurred: failed calling webhook "metallbvalidationwebhook.metallb.io": failed to call webhook: Post "https://metallb-operator-controller-manager-service.metallb-system.svc:443/validate-metallb-io-v1beta1-metallb?timeout=10s": no endpoints available for service "metallb-operator-controller-manager-service
      
      

      Steps to Reproduce:

      1. oc edit networks.operator.openshift.io cluster and set routingViaHost flag to true
        defaultNetwork:
          ovnKubernetesConfig:
            egressIPConfig: {}
            gatewayConfig:
              routingViaHost: true << This flag
      2. Wait until the update is rolled out across all nodes 
      oc get networks.operator.openshift.io cluster -o yaml
      
        - lastTransitionTime: "2023-08-31T06:22:15Z"
          status: "False"
          type: Progressing
      
      3. Try to apply sriov or metallb policy on cluster.
       Internal error occurred: failed calling webhook "metallbvalidationwebhook.metallb.io": failed to call webhook: Post "https://metallb-operator-controller-manager-service.metallb-system.svc:443/validate-metallb-io-v1beta1-metallb?timeout=10s": no endpoints available for service "metallb-operator-controller-manager-service 

      Actual results:

      local gateway mode: routingViaHost-true flag breaks connectivity to pod services from hostnetwork pods

      Expected results:

      Local gateway mode should not break connectivity of hostNetwork pods 

      Additional info:

       

            sseethar Surya Seetharaman
            nkononov@redhat.com Nikita Kononov
            Arti Sood Arti Sood
            Andrea Panattoni
            Votes:
            0 Vote for this issue
            Watchers:
            14 Start watching this issue

              Created:
              Updated:
              Resolved: