Description of problem:

idle service cannot be waked up


Version-Release number of selected component (if applicable):
4.8.0-0.nightly-2021-04-28-005607

How reproducible:
always

Steps to Reproduce:
1. oc new-project z2
  
2. create test pod 
  oc create -f https://raw.githubusercontent.com/openshift/verification-tests/master/testdata/networking/list_for_pods.json
 
  oc create -f https://raw.githubusercontent.com/openshift/verification-tests/master/testdata/networking/aosqe-pod-for-ping.json


3. # oc get svc -n z2
NAME           TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)     AGE
test-service   ClusterIP   172.30.222.181   <none>        27017/TCP   13m

4. access the service 

 oc exec -n z2 hello-pod -- curl 172.30.222.181:27017

Hello OpenShift!

5. idle service

 oc idle test-service -n z2

6. access the service again

[root@dell-per740-36 testdata]# oc exec -n z2 hello-pod -- curl --connect-timeout 10 172.30.222.181:27017
curl: (28) Connection timed out after 10000 milliseconds
command terminated with exit code 28
[root@dell-per740-36 testdata]# oc exec -n z2 hello-pod -- curl --connect-timeout 10 172.30.222.181:27017
curl: (28) Connection timed out after 10001 milliseconds
command terminated with exit code 28
  
7. # oc get ep -n z2
NAME           ENDPOINTS   AGE
test-service   <none>      20m

Actual results:

service cannot be waked up

Expected results:


Additional info:

this issue happen on all openshift-sdn and ovn-kuberntes

seems this feature cannot be worked from build 4.8.0-0.nightly-2021-04-27-231547