Bug 1961205

+++ This bug was initially created as a clone of Bug #1960401 +++

Description of problem:
Unable to access novnc url since the haproxy is configured to use the control plane ip address for the target controller instead of using it's internal API address.

# Current Failing Puddle (RHOS-16.2-RHEL-8-20210511.n.0), from /var/lib/config-data/puppet-generated/haproxy/etc/haproxy/haproxy.cfg
listen nova_novncproxy
  bind 10.0.0.119:6080 transparent
  bind 192.168.24.33:6080 transparent <-------- Using control plane ip address
  balance source
  http-request set-header X-Forwarded-Proto https if { ssl_fc }
  http-request set-header X-Forwarded-Proto http if !{ ssl_fc }
  http-request set-header X-Forwarded-Port %[dst_port]
  option tcpka
  option tcplog
  timeout tunnel 1h
  server controller-0.ctlplane.redhat.local 192.168.24.19:6080 check fall 5 inter 2000 rise 2 <------ also mapping to the control plane IP

# novnc request in haproxy logs when using control plane IP:
May 13 17:24:22 controller-0 haproxy[12]: 10.0.0.58:39130 [13/May/2021:17:24:22.037] nova_novncproxy nova_novncproxy/<NOSRV> -1/-1/0 0 SC 116/1/0/0/0 0/0

# Previous Puddle (RHOS-16.2-RHEL-8-20210420.n.0):
listen nova_novncproxy
  bind 10.0.0.133:6080 transparent
  bind 172.17.1.89:6080 transparent  <------ Uses internal api IP
  balance source
  http-request set-header X-Forwarded-Proto https if { ssl_fc }
  http-request set-header X-Forwarded-Proto http if !{ ssl_fc }
  http-request set-header X-Forwarded-Port %[dst_port]
  option tcpka
  option tcplog
  timeout tunnel 1h
  server controller-0.internalapi.redhat.local 172.17.1.65:6080 check fall 5 inter 2000 rise 2 <------ Uses internal api IP

# novnc request in haproxy when using internal api IP:
Apr 27 11:01:13 controller-0 haproxy[12]: 10.0.0.17:53700 [27/Apr/2021:11:01:12.871] nova_novncproxy nova_novncproxy/novacontrol-1.internalapi.redhat.local 1/0/394 161 -- 172/1/0/0/0 0/0

# Controller(s) with failing puddle have the novnc addressed correctly with an IP from internal api but haproxy is configured to use the control plane IP:
# IP address that the noVNC console proxy should bind to. For more information,
# refer to the documentation. (string value)
#novncproxy_host=0.0.0.0
novncproxy_host=172.17.1.136

#
# Port that the noVNC console proxy should bind to. For more information, refer
# to the documentation. (port value)
# Minimum value: 0
# Maximum value: 65535
#novncproxy_port=6080
novncproxy_port=6080

Version-Release number of selected component (if applicable):
(undercloud) [stack@undercloud-0 ~]$ cat core_puddle_version 
RHOS-16.2-RHEL-8-20210511.n.0(undercloud) 
[stack@undercloud-0 ~]$ cat /etc/rhosp-release 
Red Hat OpenStack Platform release 16.2.0 Beta (Train)


How reproducible:
100%

Steps to Reproduce:
1. Deploy a 16.2 environment with puddle RHOS-16.2-RHEL-8-20210511.n.0
2. Request a novnc console url and attempt to access the url
3.

Actual results:
Unable to access to novnc url

Expected results:
URL provided is accessible

Additional info:

# Test CI job that is failing:
https://rhos-ci-jenkins.lab.eng.tlv2.redhat.com/job/DFG-compute-nova-16.2_director-rhel-virthost-1cont_2novactl_2comp_3ceph-ipv4-geneve-ceph-tempest-composable-phase3/24/

# Test Report of Failure
https://rhos-ci-jenkins.lab.eng.tlv2.redhat.com/job/DFG-compute-nova-16.2_director-rhel-virthost-1cont_2novactl_2comp_3ceph-ipv4-geneve-ceph-tempest-composable-phase3/24/testReport/

# Environment Logs
http://rhos-ci-logs.lab.eng.tlv2.redhat.com/logs/rcj/DFG-compute-nova-16.2_director-rhel-virthost-1cont_2novactl_2comp_3ceph-ipv4-geneve-ceph-tempest-composable-phase3/24/

Did also check https://bugzilla.redhat.com/show_bug.cgi?id=1378024, but OS::TripleO::Services::NovaMetadata appears to have been missing from controller role even before this puddle.