IP of nodes being set to secondary interface IP after reboot. Checked following:

- Default gateway is set to primary interface only.
- /etc/systemd/system/kubelet.service.d/20-nodenet.conf file was absent from node.
- Execution of nodeip-configuration service failed after reboot due to a syntax error.
- "<no value>" is there in place of " $KUBELET_NODEIP_HINT".  I have checked both 00-master and 00-worker and in both machine config I could see $KUBELET_NODEIP_HINT has been replaced. This is the reason service is failing and in turn 20-nodenet.conf file is not being created and as a result kubelet choosing a random ip from available interfaces.

~~~
[arghosh@supportshell-1 03332438]$ cat 0010-must-gather.local.8237171629558956429.tar.gz/must-gather.local.8237171629558956429/quay-io-openshift-release-dev-ocp-v4-0-art-dev-sha256-8462387895747096e0aadaea8ab19129fdf19604a1a2e98698873762f73c1799/cluster-scoped-resources/machineconfiguration.openshift.io/machineconfigs/00-master.yaml |grep nodeip -B18
          ExecStart=/bin/bash -c " \
            until \
            /usr/bin/podman run --rm \
            --authfile /var/lib/kubelet/config.json \
            --net=host \
            --volume /etc/systemd/system:/etc/systemd/system:z \
            quay.io/openshift-release-dev/ocp-v4.0-art-dev@sha256:d4c84617ac0f431d42f544d39226c7e7f3e0b18088011afda69f4b67a3a2c389 \
            node-ip \
            set --retry-on-failure \
            <no value>; \
            do \
            sleep 5; \
            done"
          ExecStart=/bin/systemctl daemon-reload

          [Install]
          WantedBy=multi-user.target
        enabled: false
        name: nodeip-configuration.service
~~~ 

4.10.32

Could not reproduce by upgrading a 4.10.31 cluster to 4.10.32. Not sure if it is platform specific.

1.
2.
3.

Node IP is being set to secondary interface  without having any gateway attached to it.

IP of Primary interface having default gateway should be set  as Node's IP.

- vSphere UPI