Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-39260

traffic from external to microshift ipv6 loadbalancer svc failed

XMLWordPrintable

    • Critical
    • No
    • 3
    • uShift Sprint 259
    • 1
    • Proposed
    • False

      Description of problem:

      Setup microshift ipv6 singlestack cluster, created loadbalancer service with ETP and ITP 'cluster' on it. Sent traffic from buildvm to the external-ip, traffic was dropped. Meanwhile, sent traffic to endpoints passed, and traffic from testpod inside the cluster also passed.      

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

          4.17 

      How reproducible:

          always

      Steps to Reproduce:

          1. setup ipv6 singlestack microshift cluster 
          2. create loadbalancer svc on it.
      $ oc get svc -o yaml
      apiVersion: v1
      items:
      - apiVersion: v1
        kind: Service
        metadata:
          creationTimestamp: "2024-08-30T10:44:08Z"
          labels:
            name: test-service
          name: lbtest-etp-itp-cluster
          namespace: e2e-ushift-sdn-60968-nnnrjaz9
          resourceVersion: "103616"
          uid: 3d0fed90-0076-4fc7-b200-600f23f56dba
        spec:
          allocateLoadBalancerNodePorts: true
          clusterIP: fd02::a2e9
          clusterIPs:
          - fd02::a2e9
          externalTrafficPolicy: Cluster
          internalTrafficPolicy: Cluster
          ipFamilies:
          - IPv6
          ipFamilyPolicy: SingleStack
          ports:
          - name: http
            nodePort: 31721
            port: 27017
            protocol: TCP
            targetPort: 8080
          selector:
            name: hello-pod
          sessionAffinity: None
          type: LoadBalancer
        status:
          loadBalancer:
            ingress:
            - ip: 2620:52:0:1eb:a117:863e:a7d6:82cc
              ipMode: VIP
      kind: List
      metadata:
        resourceVersion: ""
      
      $ oc get pods -o wide
      NAME          READY   STATUS    RESTARTS   AGE   IP               NODE                   NOMINATED NODE   READINESS GATES
      hello-pod-0   1/1     Running   0          30m   fd01:0:0:1::1c   microshift-dev.local   <none>           <none>
      hello-pod-1   1/1     Running   0          29m   fd01:0:0:1::1d   microshift-dev.local   <none>           <none>
      test-pod      1/1     Running   0          29m   fd01:0:0:1::1e   microshift-dev.local   <none>           <none>
      
      $ oc get svc
      NAME                     TYPE           CLUSTER-IP   EXTERNAL-IP                         PORT(S)           AGE
      lbtest-etp-itp-cluster   LoadBalancer   fd02::cfb7   2620:52:0:1eb:a117:863e:a7d6:82cc   27017:30879/TCP   24s
      
          3. curl from buildvm (the host where the microshift installed) or node failed. Meanwhile, curl from test-pod to external ip passed.
      curl from buildvm to endpoints also passed.
      
      $ timeout 5 curl -k [2620:52:0:1eb:a117:863e:a7d6:82cc]:27017
      $ oc rsh test-pod curl -k [2620:52:0:1eb:a117:863e:a7d6:82cc]:27017
      Hello OpenShift!
      $ curl -k [fd01:0:0:1::1c]:8080
      Hello OpenShift!
      [redhat@microshift-dev ~]$ curl -k [fd01:0:0:1::1d]:8080
      Hello OpenShift!
          

      Actual results:

          traffic from external to loadbalancer svc external-ip failed.

      Expected results:

           traffic from external to loadbalancer svc external-ip should pass.

      Additional info:

          

            pacevedo@redhat.com Pablo Acevedo Montserrat
            rhn-support-yingwang Ying Wang
            Ying Wang Ying Wang
            Votes:
            0 Vote for this issue
            Watchers:
            7 Start watching this issue

              Created:
              Updated: