[ovn] [ocp 4.10.z] Service `spec.externalTrafficPolicy` does not trigger rules update in ovnkube-node pod handlers on edit, even though it does successfully update the rules if deployed explicitly with that spec value set, or if you delete the handler pods for ovn (forces a refresh).

observed in 4.10.32 and 4.10.40, tested on azure platform.

every time

1. Deploy a test pod with curlable resource in a test namespace

2. create a service from yaml exposing pod at internal clusterIP (example yaml provided by customer below)
~~~ 
apiVersion: v1
kind: Service
metadata:
  labels:
    run: test
  name: test
  annotations:
    service.beta.kubernetes.io/azure-load-balancer-internal: "true"
    service.beta.kubernetes.io/azure-load-balancer-internal-subnet: paas1
spec:
  allocateLoadBalancerNodePorts: true
  externalTrafficPolicy: Cluster ##MODIFY THIS SPEC VALUE AND OBSERVE FAIL CONDITION
  internalTrafficPolicy: Cluster
  ipFamilies:
  - IPv4
  ipFamilyPolicy: SingleStack
  ports:
  - port: 8000
    protocol: TCP
    targetPort: 8000
  selector:
    run: test
  sessionAffinity: None
  type: LoadBalancer
~~~

3. curl against service succeeds
4. edit service to change `spec.externalTrafficPolicy: local`
5. observe externalIP does not change, but healthz port updates
6. curl against same externalIP:port time out indefinitely, no response.

//workaround: 

delete service and redeploy with spec line set already to `local`, or delete ovnkube-node pod serving pod(s) to force refresh the local ruleset and allow traffic (curls subsequently will succeed).

spec change appears to update properly in the database but does not send a notification to update the ovnkube-node pod handlers (or similar) to allow traffic through once the externalTrafficPolicy spec value is changed.

spec change to service yaml should be immediately updated in DB AND update ovnkube-node handlers for same.

Attachments available and case number with specifics in next internal comment.