When Customer have scaled machines from 8 to 16 and then removed the old 8 nodes due to which the egressip did not switch to the newly scaled machines. 

Basically the issue occurs when they make new machines and get rid of the original ones. 

They just delete the label - "(k8s.ovn.org/egress-assignable: "")" that the machineset has already put on the new nodes, it remakes them, the egressip starts working. 

Problem occurs on multiple separate OpenShift clusters so this isnt a cluster specific thing. 

The steps to reproduce - 

oc project openshift-machine-api
oc edit machinesets.machine.openshift.io MACHINESET NAME <-- change replicas from 8 to 16 and altered numCPUs count up to 16 and saved
watched new machines appear, when all there
oc get egressips <-- ips on machines
oc annotate machine/NODE  -n openshift-machine-api machine.openshift.io/cluster-api-delete-machine="true" <-- for the original 8 nodes
oc edit machinesets.machine.openshift.io MACHINESET NAME <-- change replicas from 16 to 8
watched machines deleting. 
oc get egressips <-- NO ips were on any machines
oc edit node NODENAME <-- went to the  label    k8s.ovn.org/egress-assignable: "" and removed it then saved. NOTE this label was put on by the machineset. 
oc get egressips  <-- all IPs appeared on the node just edited
oc edit node NODENAME <-- went to the  label    k8s.ovn.org/egress-assignable: "" and removed it then saved for other new 7 nodes.