[Focal/Victoria/Charm-32] Unable to create LoadBalancers

[Kanshi]

Hello,

I have deployed OpenStack using MAAS+juju (Focal/Victoria) and everything works pretty well but Octavia.
Here is the scenario: I create a loadbalancer, which stays quite some time in the 'pending creation' state then turns into error.

First here the configuration that I've applied to Octavia (pretty basic):
octavia:
  openstack-origin: cloud:focal-victoria
  vip: 172.16.3.15
  os-public-hostname: octavia.example.com
  os-internal-hostname: octavia.example.com
  os-admin-hostname: octavia.example.com
  loadbalancer-topology: ACTIVE_STANDBY
  worker-multiplier: 4
  spare-pool-size: 0

I also do see that resources have been created:
$ os network list
+--------------------------------------+-------------+--------------------------------------+
| ID | Name | Subnets |
+--------------------------------------+-------------+--------------------------------------+
| eb24ed17-393a-407a-8a5f-6eccfd968890 | lb-mgmt-net | 8ce20733-c3cb-4503-adbb-7e6dd9614347 |
+--------------------------------------+-------------+--------------------------------------+
$ os network show eb24ed17-393a-407a-8a5f-6eccfd968890
+---------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------+
| Field | Value |
+---------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------+
| admin_state_up | UP |
| availability_zone_hints | |
| availability_zones | |
| created_at | 2021-03-24T00:16:04Z |
| description | |
| dns_domain | |
| id | eb24ed17-393a-407a-8a5f-6eccfd968890 |
| ipv4_address_scope | None |
| ipv6_address_scope | None |
| is_default | None |
| is_vlan_transparent | None |
| location | cloud='', project.domain_id=, project.domain_name=, project.id='79c023f31daa4db68cd0e2ca78188a5f', project.name=, region_name='RegionOne', zone= |
| mtu | 1442 |
| name | lb-mgmt-net |
| port_security_enabled | False |
| project_id | 79c023f31daa4db68cd0e2ca78188a5f |
| provider:network_type | geneve |
| provider:physical_network | None |
| provider:segmentation_id | 1001 |
| qos_policy_id | None |
| revision_number | 3 |
| router:external | Internal |
| segments | None |
| shared | False |
| status | ACTIVE |
| subnets | 8ce20733-c3cb-4503-adbb-7e6dd9614347 |
| tags | charm-octavia |
| updated_at | 2021-03-24T00:16:10Z |
+---------------------------+--------------------------------------------------------------------------------------------------------------------------------------------------+

Octavia VMs have also been created:
$ os server list --project 79c023f31daa4db68cd0e2ca78188a5f
+--------------------------------------+----------------------------------------------+--------+-----------------------------------------------------+----------------------------------------------+---------------+
| ID | Name | Status | Networks | Image | Flavor |
+--------------------------------------+----------------------------------------------+--------+-----------------------------------------------------+----------------------------------------------+---------------+
| 01bf8699-7921-4540-8292-69b56bbd505f | amphora-ed2c67d6-e6e3-476f-a159-1a1abc97888b | ACTIVE | lb-mgmt-net=fc00:6ecc:fd96:8890:f816:3eff:feba:6b53 | amphora-haproxy-x86_64-ubuntu-20.04-20210315 | charm-octavia |
| 8aa59f3f-ec2c-4a2a-9a80-eadd848efbcc | amphora-a2157d59-5bc0-4d73-947b-57a2ff1d6f14 | ACTIVE | lb-mgmt-net=fc00:6ecc:fd96:8890:f816:3eff:feca:62d8 | amphora-haproxy-x86_64-ubuntu-20.04-20210315 | charm-octavia |

But the octavia workers keeps complaining that it can't reach the VM (network unreachable):
2021-03-25 10:49:36.370 3508 WARNING octavia.amphorae.drivers.haproxy.rest_api_driver [-] Could not connect to instance. Retrying.: requests.exceptions.ConnectionError: HTTPSConnectionPool(host='fc00:6ecc:fd96:8890:f816:3eff:fe4d:1b7e', port=9443): Max retries exceeded with url: // (Caused by NewConnectionError('<urllib3.connection.VerifiedHTTPSConnection object at 0x7f233e14af10>: Failed to establish a new connection: [Errno 101] Network is unreachable'))

Which seems to be logic because the only geneve tunnel that I have on the Octavia machine doesn't seem to match the subnet:
genev_sys_6081: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 65000
        inet6 fe80::5093:f7ff:fe83:96ca prefixlen 64 scopeid 0x20<link>
        ether 52:93:f7:83:96:ca txqueuelen 1000 (Ethernet)
        RX packets 211713 bytes 17386012 (17.3 MB)
        RX errors 0 dropped 0 overruns 0 frame 0
        TX packets 231391 bytes 175493243 (175.4 MB)
        TX errors 0 dropped 67 overruns 0 carrier 0 collisions 0

and thus:
$ ping fc00:6ecc:fd96:8890::1
ping: connect: Network is unreachable

Thanks for the help!
Antoine