CSIT Sporadic failures - Elan Suite - VMs not receiving ip address

It was believed that after recent CSIT optimizations that we were not giving
enough time for the openstack instances to boot and receive their ip addresses,
but after changing our poll time to 2m (which should be more than enough) we
still see trouble.

example below, but we need to debug now

https://logs.opendaylight.org/releng/jenkins092/netvirt-csit-1node-openstack-newton-nodl-v2-upstream-stateful-carbon/791/log.html.gz#s1-s6-s1-k1