Issue:
Client performed our procedure to do a maintenance on a compute node:
https://docs.redhat.com/en/documentation/red_hat_openstack_platform/17.1/html/configuring_high_availability_for_instances/proc_performing-maintenance-compute-controller-nodes-with-instanceha_rhosp

But once the compute node came back, the private network for the instances is not working as expected. The geneve tunnels on one controller nodes are not coming back.
BFD traffic is seen in tcpdump but it stays down.

This is what we see:
Port ovn-e03d04-0
Interface ovn-e03d04-0
type: geneve
options:

{csum="true", key=flow, remote_ip="10.224.38.43", tos="0"}

bfd_status:

{diagnostic="No Diagnostic", flap_count="0", forwarding="false", remote_diagnostic="No Diagnostic", remote_state=down, state=down}

The controller node was restarted and it didn't help.

We need your help to identify why the traffic isn't working.

Attached to the case:
We enable debug for bfd on all 3 controller nodes.
We generated tcpdump from controller 1 and 2 on vla206.
We enabled debug mode for neutron on all 3 controller nodes.
We got sosreports from 3 controller + 1 compute nodes
We got openvswitch db
We got ovn dbs

Expected behavior
Network traffic should have restarted automatically

Bug impact
Traffic isn't working for that compute node

Known workaround
None