Wrong listen-address for dnsmasq service after Updating to 9.1

There are two ways how to update Fuel from 9.0 to 9.1:
1) Install 9.0 (fuel-devops framework with test group prepare_release could be used)
2) Create a new yum repo file with appropriate repository (for ex. [1])
3) Run yum update
4) Run `fuel-bootstrap build --verbose --debug --activate --label 'new-bootstrap'` on the master ndoe to rebuild bootstrap.

Result: on the each step I see right ip address at the /etc/dnsmasq.d/default.conf, it is the master node ip address, for me it is 10.109.0.2

The second way:
1) Export an environment variables UPDATE_MASTER, UPDATE_FUEL_MIRROR, EXTRA_DEB_REPOS, MIRROR_UBUNTU (you can see my deploy.sh script in attachments)
2) Deploy an environment (group prepare_release is used again)

In this way we also run yum update and a bootstrap rebuild under tests.

Result: as I see at the dnsmasq.conf listen-address is wrong:

[root@nailgun ~]# cat /etc/dnsmasq.d/default.conf

listen-address=172.17.0.1
dhcp-range=default,10.109.0.3,10.109.0.254,255.255.255.0,120m
dhcp-option=net:default,option:router,10.109.0.1
dhcp-boot=net:default,pxelinux.0,boothost,10.109.0.2
dhcp-match=set:ipxe,175
dhcp-option-force=tag:ipxe,210,http://10.109.0.2/cobbler/boot/

This ip address belongs to docker0 (?!) interface. After some manipulations this interface has disappeared from the system.

And as a result:

[root@nailgun ~]# systemctl status dnsmasq.service
● dnsmasq.service - DNS caching server.
   Loaded: loaded (/usr/lib/systemd/system/dnsmasq.service; enabled; vendor preset: disabled)
  Drop-In: /etc/systemd/system/dnsmasq.service.d
           └─restart.conf
   Active: activating (auto-restart) (Result: exit-code) since Tue 2016-08-23 13:32:52 UTC; 1s ago
  Process: 3498 ExecStart=/usr/sbin/dnsmasq -k (code=exited, status=2)
 Main PID: 3498 (code=exited, status=2)

Aug 23 13:32:52 nailgun.test.domain.local systemd[1]: Child 3498 belongs to dnsmasq.service
Aug 23 13:32:52 nailgun.test.domain.local systemd[1]: dnsmasq.service: main process exited, code=exited, status=2/INVALIDARGUMENT
Aug 23 13:32:52 nailgun.test.domain.local systemd[1]: dnsmasq.service changed running -> failed
Aug 23 13:32:52 nailgun.test.domain.local systemd[1]: Unit dnsmasq.service entered failed state.
Aug 23 13:32:52 nailgun.test.domain.local systemd[1]: dnsmasq.service failed.
Aug 23 13:32:52 nailgun.test.domain.local systemd[1]: dnsmasq.service changed failed -> auto-restart
Aug 23 13:32:52 nailgun.test.domain.local systemd[1]: dnsmasq.service: cgroup is empty
Aug 23 13:32:52 nailgun.test.domain.local dnsmasq[3498]: dnsmasq: failed to create listening socket for 172.17.0.1: Cannot assign requested address

[root@nailgun ~]# netstat -nap| grep masq
[root@nailgun ~]#

If I tried to boot some node, it couldn't obtain an ip address and boot from pxe.
So, I think that we need check that config for a valid ip address and re-configure it if it needed.
Or to bind to the interface.

[1] http://mirror.fuel-infra.org/mos-repos/centos/mos9.0-centos7/snapshots/proposed-latest/x86_64/