service configured with a nodeport can't be reached until after restart of ovnkube-master

After a clean install of MicroShift from ISO, a service configured with a nodeport can not be reached

[redhat@localhost ~]$ rpm -qi microshift
Name        : microshift
Version     : 4.12.0_20230119163440_3e0c7c14
Release     : 1.el8
Architecture: x86_64
Install Date: Thu Jan 19 13:20:00 2023
Group       : Unspecified
Size        : 118331948
License     : ASL 2.0
Signature   : (none)
Source RPM  : microshift-4.12.0_20230119163440_3e0c7c14-1.el8.src.rpm
Build Date  : Thu Jan 19 13:13:51 2023
Build Host  : microshift-dev
Relocations : (not relocatable)
URL         : https://github.com/openshift/microshift
Summary     : MicroShift service
Description :
The microshift package provides an OpenShift Kubernetes distribution optimized for small form factor and edge computing.

Always

1. Create a new microshift cluster (https://mastern-jenkins-csb-openshift-qe.apps.ocp-c1.prod.psi.redhat.com/job/microshift/job/microshift-deploy-iso/build?delay=0sec)
2. export KUBECONFIG=kubeconfig
3. oc create ns test
4. oc create -n test -f ./nodeport_test_pod.yaml
5. oc create -n test -f ./nodeport_test_service.yaml
6. oc get service -n test
NAME        TYPE       CLUSTER-IP     EXTERNAL-IP   PORT(S)           AGE
hello-pod   NodePort   10.43.19.147   <none>        27017:32170/TCP   5m31s
*NOTE PORT NUMBERS (e.g. 32170)
7. oc get node
NOTE NODE NAME (e.g. dhcp-1-235-100.arm.eng.rdu2.redhat.com)
8 curl NODE:PORT
Expect "Customer-Red Test Nodeport" response  

$ oc get all -n test
NAME                  READY   STATUS    RESTARTS   AGE
pod/hello-pod-hkqpp   1/1     Running   0          12m

NAME                              DESIRED   CURRENT   READY   AGE
replicationcontroller/hello-pod   1         1         1       12m

NAME                TYPE       CLUSTER-IP     EXTERNAL-IP   PORT(S)           AGE
service/hello-pod   NodePort   10.43.19.147   <none>        27017:32170/TCP   12m

$ oc get node
NAME                                     STATUS   ROLES                         AGE   VERSION
dhcp-1-235-100.arm.eng.rdu2.redhat.com   Ready    control-plane,master,worker   8h    v1.25.0

$ curl dhcp-1-235-100.arm.eng.rdu2.redhat.com:32170
curl: (7) Failed to connect to dhcp-1-235-100.arm.eng.rdu2.redhat.com port 32170: No route to host

$ sudo iptables-save | grep 32170
-A OVN-KUBE-NODEPORT -p tcp -m addrtype --dst-type LOCAL -m tcp --dport 32170 -j DNAT --to-destination 10.43.19.147:27017   

$ curl dhcp-1-235-100.arm.eng.rdu2.redhat.com:32170
Customer-Red Test Nodeport

$ Workaround requires restart of the ovnkube-master pod
(e.g. oc delete pod -n openshift-ovn-kubernetes ovnkube-master-d598f)