This is a clone of issue OCPBUGS-43740. The following is the description of the original issue:
—
After changing internalJoinSubnet,internalTransitSwitchSubnet, on day2 and do live migration.  ovnkube node pod crashed

network part as below the service cidr has same subnet with the ovn default internalTransitSwitchSubnet

    clusterNetwork:
    - cidr: 100.64.0.0/15
      hostPrefix: 23
    serviceNetwork:
    - 100.88.0.0/16

and then:

oc patch network.operator.openshift.io cluster --type='merge' -p='{"spec":{"defaultNetwork":{"ovnKubernetesConfig":{"ipv4":{"internalJoinSubnet": "100.82.0.0/16"}}}}}'
oc patch network.operator.openshift.io cluster --type='merge' -p='{"spec":{"defaultNetwork":{"ovnKubernetesConfig":{"ipv4":{"internalTransitSwitchSubnet": "100.69.0.0/16"}}}}}'

with error: 

 start-ovnkube-node ${OVN_KUBE_LOG_LEVEL} 29103 29105 State: Waiting Reason: CrashLoopBackOff Last State: Terminated Reason: Error Message: EmptyLbEvents:false PodIP: RawNoHostSubnetNodes:migration.network.openshift.io/plugin= NoHostSubnetNodes:<nil> HostNetworkNamespace:openshift-host-network PlatformType:AWS HealthzBindAddress:0.0.0.0:10256 CompatMetricsBindAddress: CompatOVNMetricsBindAddress: CompatMetricsEnablePprof:false DNSServiceNamespace:openshift-dns DNSServiceName:dns-default} Metrics:{BindAddress: OVNMetricsBindAddress: ExportOVSMetrics:false EnablePprof:false NodeServerPrivKey: NodeServerCert: EnableConfigDuration:false EnableScaleMetrics:false} OvnNorth:{Address: PrivKey: Cert: CACert: CertCommonName: Scheme: ElectionTimer:0 northbound:false exec:<nil>} OvnSouth:{Address: PrivKey: Cert: CACert: CertCommonName: Scheme: ElectionTimer:0 northbound:false exec:<nil>} Gateway:{Mode:shared Interface: EgressGWInterface: NextHop: VLANID:0 NodeportEnable:true DisableSNATMultipleGWs:false V4JoinSubnet:100.64.0.0/16 V6JoinSubnet:fd98::/64 V4MasqueradeSubnet:100.254.0.0/17 V6MasqueradeSubnet:fd69::/125 MasqueradeIPs:

DisablePacke