When the node has multiple ports, baremetal-runtimecfg is smart and will use the IP on eth0 for NODE_IP used by kubelet. But since https://github.com/openshift/machine-config-operator/pull/4388 - we are now forcing it to be 0.0.0.0.     

4.18, 4.17, 4.16    

Install a cluster (standalone or hosted) where workers (or nodepools) have multiple ports (and addresses). Kubelet will be wrongly configure to listen on 0.0.0.0:

/usr/bin/kubelet --config=/etc/kubernetes/kubelet.conf --bootstrap-kubeconfig=/etc/kubernetes/kubeconfig --kubeconfig=/var/lib/kubelet/kubeconfig --container-runtime-endpoint=/var/run/crio/crio.sock --runtime-cgroups=/system.slice/crio.service --node-labels=node-role.kubernetes.io/worker,node.openshift.io/os_id=rhcos, --node-ip=0.0.0.0 --address=0.0.0.0 --minimum-container-ttl-duration=6m0s --volume-plugin-dir=/etc/kubernetes/kubelet-plugins/volume/exec --cloud-provider=external --hostname-override=emacchi-hcp-sriov-msrxd --pod-infra-container-image=registry.ci.openshift.org/ocp/4.17-2024-08-14-124358@sha256:3152b2e42d8bb97baa03cf566b85914d7d6b420d7112611c3e44f6edf76de0dd --system-reserved=cpu=500m,memory=1Gi,ephemeral-storage=1Gi --v=2    

Kubelet listening on 0.0.0.0 instead of its primary IP.    

Kubelet should be listening on its primary IP found by baremetal-runtimecfg: https://github.com/openshift/baremetal-runtimecfg/blob/8371408693c86490b7b4d207aa6a732d15137757/cmd/runtimecfg/node-ip.go#L141