After scale down the default ingresscontroller replicas to 0 and remove all default router pods, there is still one keepalived pod in MASTER state and holding the ingress Virtual IP.

Comparing the ipfailover (https://docs.openshift.com/container-platform/4.10/networking/configuring-ipfailover.html) which using the same image, if the track_script failed then it enter FAULT state.  

4.10.58

always

1. scale down the replicas of default ingresscontroller to zero 
$ oc -n openshift-ingress-operator scale ingresscontroller/default --replicas=0

$ oc -n openshift-ingress get pod
No resources found in openshift-ingress namespace.


2. check the keepalivd pods running on worker nodes and their logs 
$ oc -n openshift-vsphere-infra get pod -l app=vsphere-infra-vrrp
NAME                                         READY   STATUS    RESTARTS   AGE
keepalived-hongli-vs410-k6n7k-master-0       2/2     Running   0          7h50m
keepalived-hongli-vs410-k6n7k-master-1       2/2     Running   0          7h51m
keepalived-hongli-vs410-k6n7k-master-2       2/2     Running   0          7h50m
keepalived-hongli-vs410-k6n7k-worker-2mvxd   2/2     Running   0          7h42m
keepalived-hongli-vs410-k6n7k-worker-46sjv   2/2     Running   0          7h42m
keepalived-hongli-vs410-k6n7k-worker-ddmtg   2/2     Running   0          7h42m

$ oc -n openshift-vsphere-infra logs keepalived-hongli-vs410-k6n7k-worker-ddmtg
<......>
Fri Apr 28 04:57:50 2023: (hongli-vs410_INGRESS) Received advert from 172.31.249.204 with lower priority 20, ours 20, forcing new election
Fri Apr 28 04:57:50 2023: (hongli-vs410_INGRESS) Sending/queueing gratuitous ARPs on ens192 for 172.31.248.53
Fri Apr 28 04:57:50 2023: Sending gratuitous ARP on ens192 for 172.31.248.53
Fri Apr 28 04:57:50 2023: Sending gratuitous ARP on ens192 for 172.31.248.53


$ oc -n openshift-vsphere-infra rsh keepalived-hongli-vs410-k6n7k-worker-ddmtg
Defaulted container "keepalived" out of: keepalived, keepalived-monitor, render-config-keepalived (init)
sh-4.4# ip a | grep 172
    inet 172.31.249.245/23 brd 172.31.249.255 scope global dynamic noprefixroute ens192
    inet 172.31.248.53/32 scope global ens192                <<<---ingress vip
sh-4.4# 

$ oc get infrastructures.config.openshift.io cluster -oyaml
<......>
status:
  platform: VSphere
  platformStatus:
    type: VSphere
    vsphere:
      apiServerInternalIP: 172.31.248.49
      ingressIP: 172.31.248.53


3.

all keepalived pods on worker nodes have the same priority 20, and the one with highest IP win the election and takes MASTER state.

if no default router pod running, the keepalived should be in FAULT state so it won't be able to enter MASTER state. 

the keepalived.conf looks like 

# TODO: Improve this check. The port is assumed to be alive.
# Need to assess what is the ramification if the port is not there.
vrrp_script chk_ingress {
    script "/usr/bin/timeout 0.9 /usr/bin/curl -o /dev/null -Lfs http://localhost:1936/healthz/ready"
    interval 1
    weight 20
    rise 3
    fall 2
}vrrp_script chk_default_ingress {
    script "/usr/bin/timeout 4.9 /etc/keepalived/chk_default_ingress.sh"
    interval 5
    weight 50
    rise 3
    fall 2
}

vrrp_instance hongli-vs410_INGRESS {
    state BACKUP
    interface ens192
    virtual_router_id 2
    priority 20
    advert_int 1
    
    authentication {
        auth_type PASS
        auth_pass hongli-vs410_ingress_vip
    }
    virtual_ipaddress {
        172.31.248.53/32
    }
    track_script {
        chk_ingress
        chk_default_ingress
    }
}