On the overcloud nodes deployed by instack-virt-setup, /etc/resolv.conf looks like this: search openstacklocal nameserver 192.168.122.1 That's not a useful address for either of these nodes, on which external connectivity -- at least on the controller -- is via eth0/br-ex on the 192.0.2.0/24 network. Even worse, on the controller "192.168.122.1" is actually the address of the local "virbr0" interface configured by libvirt, so DNS requests aren't going to go anywhere useful. There are obviously a number of ways to fix this, including setting up dnsmasq on the undercloud node and using that instead. How is this supposed to work? Did the deployments script screw up, or did I miss something in the docs? -- Lars Kellogg-Stedman <lars(a)redhat.com&gt; | larsks @ {freenode,twitter,github} Cloud Engineering / OpenStack | http://blog.oddbit.com/