Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-59657

pod(VM) on default network->pod on localnet on same node is broken

XMLWordPrintable

    • Quality / Stability / Reliability
    • False
    • Hide

      None

      Show
      None
    • 5
    • None
    • None
    • None
    • None
    • CORENET Sprint 274, CORENET Sprint 275, CORENET Sprint 276
    • 3
    • Customer Escalated
    • In Progress
    • Bug Fix
    • Hide
      Previously, a pod with a secondary interface in an OVN-Kubernetes Localnet network (mapped to the br-ex bridge) could communicate with pods on the same node that used the default network for connectivity only if the Localnet IP addresses were within the same subnet as the host network. With this release, the localnet IP addresses can be drawn from any subnet; in this generalized case, an external router outside the cluster is expected to connect the localnet subnet to the host network.
      Show
      Previously, a pod with a secondary interface in an OVN-Kubernetes Localnet network (mapped to the br-ex bridge) could communicate with pods on the same node that used the default network for connectivity only if the Localnet IP addresses were within the same subnet as the host network. With this release, the localnet IP addresses can be drawn from any subnet; in this generalized case, an external router outside the cluster is expected to connect the localnet subnet to the host network.
    • None
    • None
    • None
    • None

      Description of problem:

      Customer has upgraded to 4.18.19 and still states they see the issue originally described in this bug. I'm leaving the remaining info from the original bug (minus the releases) 

       

      A VM with an IP address in a OVN localnet network is unreachable from other VMs/Pods running in the same node and using the pod network. VMs/Pods running in different nodes work fine.

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

      OCP 4.18.19
      CNV 4.18.8
      

      How reproducible:

      100%

      Steps to Reproduce:

      1. Create a localnet mapping and net-attach-def:
      
      ```
      apiVersion: nmstate.io/v1
      kind: NodeNetworkConfigurationPolicy
      metadata:
        name: localnet-mappings
      spec:
        desiredState:
          ovn:
            bridge-mappings:
            - bridge: br-ex
              localnet: ovn-localnet-1
              state: present
      ```
      
      ```
      apiVersion: k8s.cni.cncf.io/v1
      kind: NetworkAttachmentDefinition
      metadata:
        name: localnet-network
        namespace: test
      spec:
        config: |
          {
                  "cniVersion": "0.3.1",
                  "name": "ovn-localnet-1",
                  "type": "ovn-k8s-cni-overlay",
                  "topology": "localnet",
                  "netAttachDefName": "test/localnet-network"
          }
      ```
      
      2. Create a target VM with its only NIC using the test/localnet-network NAD
      3. Create a test VM (or pod) with its only NIC in the pod network
      4. Ping from the test VM to the target VM. Compare the results from both VMs running in the same node or in different nodes.

      Actual results:

      The ping from the test VM to the target VM only works if they are in different nodes.

      Expected results:

      Ping always working.

      Additional info:

       

              rravaiol@redhat.com Riccardo Ravaioli
              jevan2 Jonathan Evan
              None
              None
              Arti Sood Arti Sood
              None
              Votes:
              0 Vote for this issue
              Watchers:
              12 Start watching this issue

                Created:
                Updated: