Description of problem:

This problem caused a network outage on 3 compute nodes when performing an undercloud update (it triggers an os-update-config on the overcloud)

# Created by cloud-init on instance boot automatically, do not edit.
#
BOOTPROTO=dhcp
DEVICE=em1
HWADDR=XX:XX:XX:XX:XX:c0
ONBOOT=yes
TYPE=Ethernet
USERCTL=no

em1 should be part of a ovs-bond.

Whe os-update-config was triggered the following command failed and all the interfaces configuration got updated with safe default, causing a connectivity outage:

Oct  9 17:18:33 XXXX os-collect-config: [2019/10/09 05:18:33 PM] [INFO] Running ovs-appctl bond/set-active-slave ('bond1', 'em1')
...
...
Oct  9 17:18:33 XXXX os-collect-config: oslo_concurrency.processutils.ProcessExecutionError: Unexpected error while running command.
Oct  9 17:18:33 XXXX os-collect-config: Command: /bin/ovs-appctl bond/set-active-slave bond1 em1
Oct  9 17:18:33 XXXX os-collect-config: Exit code: 2

Here we can see cloud-init logs:

2019-10-10 15:35:09,019 - cloud_config.py[DEBUG]: Merging by applying [('dict', ['replace']), ('list', []), ('str', [])]
2019-10-10 15:35:09,060 - cloud_config.py[DEBUG]: Merging by applying [('dict', ['replace']), ('list', []), ('str', [])]
2019-10-10 15:35:09,064 - handlers.py[DEBUG]: finish: init-network/consume-user-data: SUCCESS: reading and applying user-data
2019-10-10 15:35:09,064 - handlers.py[DEBUG]: start: init-network/consume-vendor-data: reading and applying vendor-data
2019-10-10 15:35:09,064 - handlers.py[DEBUG]: finish: init-network/consume-vendor-data: SUCCESS: reading and applying vendor-data


2019-10-11 08:45:55,088 - stages.py[DEBUG]: applying net config names for {'version': 1, 'config': [{'subnets': [{'type': 'dhcp'}], 'type': 'physical', 'name': 'em1', 'mac_address': '14:18:77:60:39:c0'}]}
<=========

2019-10-11 08:46:56,109 - stages.py[DEBUG]: applying net config names for {'version': 1, 'config': [{'subnets': [{'type': 'dhcp'}], 'type': 'physical', 'name': 'br-ex', 'mac_address': '14:18:77:60:39:c0'}]}
2019-10-11 08:46:56,112 - stages.py[WARNING]: Failed to rename devices: duplicate mac found! both 'em1' and 'br-ex' have mac '14:XX:XX:60:XX:c0'
<====

2019-10-11 08:46:56,146 - handlers.py[DEBUG]: start: init-network/consume-user-data: reading and applying user-data
2019-10-11 08:46:56,149 - cloud_config.py[DEBUG]: Merging by applying [('dict', ['replace']), ('list', []), ('str', [])]
2019-10-11 08:46:56,191 - cloud_config.py[DEBUG]: Merging by applying [('dict', ['replace']), ('list', []), ('str', [])]
2019-10-11 08:46:56,195 - handlers.py[DEBUG]: finish: init-network/consume-user-data: SUCCESS: reading and applying user-data
2019-10-11 08:46:56,195 - handlers.py[DEBUG]: start: init-network/consume-vendor-data: reading and applying vendor-data
2019-10-11 08:46:56,195 - handlers.py[DEBUG]: finish: init-network/consume-vendor-data: SUCCESS: reading and applying vendor-data


I can't see why cloud-init is having this behaviour

Version-Release number of selected component (if applicable):
OSP10

How reproducible:
each boot

Steps to Reproduce:
1. reboot
2.
3.

Actual results:
cloud-init apply the safe defaults and is not in sync with os-net-config

Expected results:
cloud-init must not overwrite net interfaces configuration

Additional info: