tripleo-ci-centos-7-containerized-undercloud-upgrades fails restarting xinetd

[James Slagle]

tripleo-ci-centos-7-containerized-undercloud-upgrades failing with:

2018-10-05 02:21:01 | TASK [restart xinetd in order to free tftp port] *******************************
2018-10-05 02:21:02 | fatal: [centos-7-rax-dfw-0002630889]: FAILED! => {"changed": false, "msg": "Could not find the requested service xinetd: host"}
2018-10-05 02:21:02 |
2018-10-05 02:21:02 | PLAY RECAP *********************************************************************
2018-10-05 02:21:02 | centos-7-rax-dfw-0002630889 : ok=40 changed=16 unreachable=0 failed=1

Failed job log:
http://logs.openstack.org/01/607701/1/check/tripleo-ci-centos-7-containerized-undercloud-upgrades/64dac96/

Appears to be b/c Rocky templates are used during the upgrade.

Patch on master is:
https://review.openstack.org/#/c/607966/

Once that is backported to rocky, the issue would be "fixed", but we still shouldn't be using the ROcky templates to upgrade to master.