Description of the problem:

When upgrading SNO installed via assisted-installer to the EC version of 4.13, the local dnsmasq process in the node is not listening on all interfaces, and only listens for localhost loopback.

It makes kubelet and kube-apiserver unable to resolve the fqdn and api/api-int by locally requesting dns resolution from the dnsmasq process.

How reproducible:

100%

Steps to reproduce:

1. Upgrade to registry.ci.openshift.org/rhcos-devel/ocp-4.13-9.0:4.13.0-ec.1

Actual results:

SNO upgrade will stuck and does not proceed.

Expected results:

Successful SNO upgrade.