Bug 2048841

This is a follow up from bug 2029742.

When new pods that are expected to have egressip nat and lrpolicies are
created while ovn-k master is down, they remain missing, even after
ovn-k8 is started.

Steps to Reproduce:
1. Label one node as egress node
2. Create egressip object
 oc get egressip -o yaml
....
  spec:
    egressIPs:
    - 172.31.249.117
    namespaceSelector:
      matchLabels:
        org: pm
    podSelector: {}
  status:
    items:
    - egressIP: 172.31.249.117
      node: compute-0
....
3. Create ns ds36l and 10 pods in it, label org=pm  to the namespace
4. scale the CNO to 0 
 oc scale deployment network-operator -n openshift-network-operator --replicas 0
5.Delete ovnkube-master ds
Scale test pods replicas to 20
6. scale the CNO to 1
oc scale deployment network-operator -n openshift-network-operator --replicas 1
deployment.apps/network-operator scaled

7. Check  lr-policy-list and snat
 ovn-nbctl lr-policy-list ovn_cluster_router | grep "100 "

8. Nat rules are also not correct. They should have been added for the new
pods that were started while ovn-k8 was down. 

sh-4.4# ovn-nbctl --format=csv --no-heading find nat external_ids:name=egressip

Actual results:
lr-policy-list  and snat rules are not added for the new pod instances.


Expected results:
No stale lr-policy-list  and snat rules, and accounting for the all the ones running.