1. Proposed title of this feature request
RFE for customization for nodeIP in kubelet.service

2. What is the nature and description of the request?
When all nodes in a cluster has two NIC attached, there should be an way to set Which NIC to use during installation. nodeip-configuration.service service runs a container to choose the IP and write it to /etc/systemd/system/kubelet.service.d/20-nodenet.conf file and the kubelet.service reads KUBELET_NODE_IP from /etc/systemd/system/kubelet.service.d/20-nodenet.conf file. When I run below command I see the IP of NIC which is configured as the default gateway is always chosen as per [1].

1. podman run -it --net=host quay.io/openshift-release-dev/ocp-v4.0-art-dev@sha256:c0e6c751d3f42c280624f88bc91e1e27fa602a09fa57a41d0c34095d853e9e98 node-ip show

[1] - https://github.com/openshift/sdn/blob/7106dab9c5785ae163237d443de0ddc19cb1f040/vendor/k8s.io/apimachinery/pkg/util/net/interface.go#L395-L398

3. Why does the customer need this? (List the business requirements here)
In our environment, there are multiple network segments serving different purposes due to security reasons. We need to configure multiple NICs in OCP with default route set to the one not used to deploy OCP cluster to achieve traffic routing out from OCP clusters.
Apart from that, we also face inconsistent network device naming issue as the cluster is deployed in VMware. Hence, when there is more than one NICs declared to RHCOS, the machine-config-operator points to IP from the NIC that is first in sequence. I.E. ens192 is the NICs used to deploy OCP but the machine-config-operator tries to curl to ip declared in NICs ens161.

4. List any affected packages or components.