rsyncd started by both systemd and xinetd

Detailed bug description:
 On fuel master in version 9.1 both systemd and xinetd are trying to start rsyncd daemon.
 It leads to situation when one of this action will fail since rsyncd daemon will already listen on the port.
 Example messages from journalctl -xe:

   Feb 03 13:51:42 myhostname rsyncd[14029]: rsyncd version 3.1.0 starting, listening on port 873
   Feb 03 13:51:42 myhostname rsyncd[14029]: bind() failed: Address already in use (address-family 2)
   Feb 03 13:51:42 myhostname rsyncd[14029]: unable to bind any inbound sockets on port 873
   Feb 03 13:51:42 myhostname rsyncd[14029]: rsync error: error in socket IO (code 10) at socket.c(555) [Receiver=3.1.0]
   Feb 03 13:51:42 myhostname systemd[1]: Received SIGCHLD from PID 14029 (rsync).
   Feb 03 13:51:42 myhostname systemd[1]: Child 14029 (rsync) died (code=exited, status=10/n/a)
   Feb 03 13:51:42 myhostname systemd[1]: Child 14029 belongs to rsyncd.service
   Feb 03 13:51:42 myhostname systemd[1]: rsyncd.service: main process exited, code=exited, status=10/n/a
   Feb 03 13:51:42 myhostname systemd[1]: rsyncd.service changed running -> failed
   Feb 03 13:51:42 myhostname systemd[1]: Unit rsyncd.service entered failed state.
   Feb 03 13:51:42 myhostname systemd[1]: rsyncd.service failed.

Port is already in use by rsyncd instance started by xinetd:

netstat -lpen --tcp | grep 873
tcp 0 0 10.20.0.2:873 0.0.0.0:* LISTEN 0 21297 1142/xinetd

Steps to reproduce:
 Instal Fuel 9.0
 Upgrade to Fuel 9.1 using this doc: https://docs.mirantis.com/openstack/fuel/fuel-9.1/release-notes/update-product/update-master-9-9-1.html
 Reboot fuel master after upgrade.
Expected results:
 rsyncd should be started only once.
Actual result:
 Both systemd and xinted are trying to start rsyncd
Workaround:
 We can disable systemd rsyncd.service:
   systemctl disable rsyncd.service
Impact:
 Beside error in logs it should not break deployment.
Additional information:
 Are we still prefer xinetd over systemd for starting rsyncd?