The ovn localnet interface configured by nmstate nncp does not appear to be removed successfully. This seems happening more on nodes where there are pods with net-attach-def referencing the OVN localnet interface. Once the applied configuration is changed to absent status, the nnce resources change the status to failedtoconfigure and report the following messages:


      WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      INFO  nmstate::nm::show] Got unsupported interface type generic: genev_sys_6081,
      ignoring\n[2024-08-19T12:18:04Z ERROR nmstate::query_apply::inter_ifaces] VerificationError:
      Absent/Down interface bridge-name/ovs-bridge still found as OvsBridge(OvsBridgeInterface
      { base: BaseInterface { name: \"bridge-name\", profile_name: None, description:
 

    OCP 4.15 

    Easily

1. Apply the following configuration to the platform: 


# oc get nncp  ovs-localnet-rep -oyaml
apiVersion: nmstate.io/v1
kind: NodeNetworkConfigurationPolicy
metadata:
  annotations:
  name: ovs-localnet-rep
spec:
  desiredState:
    interfaces:
    - bridge:
        allow-extra-patch-ports: true
        options:
          stp: true
        port:
        - name: bond1
      ipv4:
        enabled: false
      name: newbr
      state: up
      type: ovs-bridge
    ovn:
      bridge-mappings:
      - bridge: newbr
        localnet: nad-test
        state: present
  nodeSelector:
    nmstate: test

# oc get nncp  bondreproducer -oyaml
apiVersion: nmstate.io/v1
kind: NodeNetworkConfigurationPolicy
metadata:
  annotations:
  name: bondreproducer
spec:
  desiredState:
    interfaces:
    - link-aggregation:
        mode: active-backup
        port:
        - ensxxx
        - ensxxx
      name: bond1
      state: up
      type: bond
  nodeSelector:
    nmstate: test

2. Check if the configuration is going to be applieds to the respective nodes: 

# oc get nnce
NAME                                                          STATUS      STATUS AGE   REASON
master1.bm-upi.<test>.bondreproducer   Available   31m          SuccessfullyConfigured
master1.bm-upi.<test>.ovs-localnet-rep     Available   29m          SuccessfullyConfigured
master2.bm-upi.<test>.bondreproducer   Available   87m          SuccessfullyConfigured
master2.bm-upi.<test>.ovs-localnet-rep     Available   29m          SuccessfullyConfigured

3. Change the nncp configuration from bridge and localnet interface to absent: 


 # oc get nncp  ovs-localnet-rep -oyaml
apiVersion: nmstate.io/v1
kind: NodeNetworkConfigurationPolicy
metadata:
  annotations:
  name: ovs-localnet-rep
spec:
  desiredState:
    interfaces:
    - bridge:
        allow-extra-patch-ports: true
        options:
          stp: true
        port:
        - name: bond1
      ipv4:
        enabled: false
      name: newbr
      state: absent <------
      type: ovs-bridge
    ovn:
      bridge-mappings:
      - bridge: newbr
        localnet: nad-test
        state: absent <-----
  nodeSelector:
    nmstate: test

 
4. Confirm that the nnce changed the status to failedtoconfigure:

# oc get nnce master2.bm-upi.brunolabocp.local.ovs-localnet-rep03877931
NAME                                                        STATUS    STATUS AGE   REASON
master2.bm-upi.<test>.ovs-localnet-rep   Failing   33s          FailedToConfigure

5. Confirm that there is any pod referencing the OVN localnet interface by a net-attach-def resource. The issue seems happening more on node where there are pods with net-attach-def referenced: 

$ oc get net-attach-def -oyaml 

    config: |-
      {
        "cniVersion": "0.3.1",
        "name": "nad-test",
        "type": "ovn-k8s-cni-overlay",
        "topology":"localnet",
        "subnets": "xx.xx.xx.0/24",
        "mtu": 1300,
        "netAttachDefName": "test-nad/nad-test"
      }

$ oc get pod -oyaml 

apiVersion: v1
kind: Pod
metadata:
  annotations:
    k8s.v1.cni.cncf.io/network-status: |-
      [{
          "name": "test-nad/nad-test",
          "interface": "net1",
          "ips": [
              "xx.xx.xx.51"
          ],
          "mac": "xx:xx:xx:xx:xx:xx",
          "dns": {}
      }]

6. If the bridge is removed from OVS and the configuration is reapplied, the issue is resolved: 

$ oc debug node/<node-name>
$ chroot /host
sh-5.1# ovs-vsctl list-br <--- find the bridge in the configuration
sh-5.1# ovs-vsctl del-br <bridge-name> <---- Delete the bridge

7. Force a nmstate reapply configuration

    The following issue is reported by the nnce resource and the configuration is not applied: 

    WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      WARN  nmstate::ovsdb::show] Unknown OVS interface type \n[2024-08-19T12:18:04Z
      INFO  nmstate::nm::show] Got unsupported interface type generic: genev_sys_6081,
      ignoring\n[2024-08-19T12:18:04Z ERROR nmstate::query_apply::inter_ifaces] VerificationError:
      Absent/Down interface bridge-name/ovs-bridge still found as OvsBridge(OvsBridgeInterface

    It is expected that the configuration finishes successfully and the interface is removed. When this issue is combined with the bug RHEL-40683, the configuration is reapplied every time that the node label is changed. This often causes failedtoconfigure in the platform.