Mirror host ping check fails

[Clark Boylan]

We have jobs that fail the mirror host ping check. This is run early in the job to fail quickly if we don't expect the host networking to otherwise function and fail in obtuse ways. However, we aren't currently tracking when this occurs. Add this bug to track it with e-r.

Example from http://logs.openstack.org/10/466910/1/gate/gate-grenade-dsvm-neutron-ubuntu-xenial/6be4506/logs/devstack-gate-setup-host.txt.gz

2017-05-23 21:50:20.418 | localhost | SUCCESS | network_sanity_check : Get NODEPOOL_MIRROR_HOST >>
mirror.la1.citycloud.openstack.org
2017-05-23 21:50:20.627 | localhost | SUCCESS | network_sanity_check : Get NODEPOOL_PYPI_MIRROR >>
http://mirror.la1.citycloud.openstack.org/pypi/simple
2017-05-23 21:50:30.888 | localhost | FAILED | network_sanity_check : Perform ping check | rc=1 >>
PING mirror.la1.citycloud.openstack.org (89.46.82.21) 56(84) bytes of data.

--- mirror.la1.citycloud.openstack.org ping statistics ---
1 packets transmitted, 0 received, 100% packet loss, time 0ms
 to retry, use: --limit @/home/jenkins/workspace/gate-grenade-dsvm-neutron-ubuntu-xenial/devstack-gate/playbooks/setup_host.retry