My environment uses:
Kolla version kolla-ansible 18.8.1.dev13
Openstack version: 2024.1 Caracal
About the migration: first I stopped Neutron and cleaned everything I needed to prepare the base for OVN. After that, I deployed the OVN plugin, and to finish, I deployed Neutron again with the OVN plugin along with some configurations in ML2. The migration only worked after these ML2 configurations; then I restarted the Neutron containers and it worked.
After the migration the openstack host worked; I could do everything — ping other VMs, ping the internet, assign floating IPs — but if I restarted my openstack host, the bug occurred. I will send some logs and database records for you to take a look.
To fix the bug I need to go into the MariaDB table and delete the entry; after that it works. But I’m not sure if this is a Kolla Ansible bug or if I’m doing something wrong.
this is before the migration.
MariaDB [neutron]> select * from providerresourceassociations;
+---------------+--------------------------------------+
| provider_name | resource_id |
+---------------+--------------------------------------+
| single_node | a186f98b-eb36-4baf-a753-e6178681eedb |
+---------------+--------------------------------------+
after the migration stay the same.
and after i restart the host or restart the neutron_server this happened.
2025-12-11 16:40:23.668 41 ERROR futurist.periodics neutron_lib.objects.exceptions.NeutronDbObjectDuplicateEntry: Failed to create a duplicate ProviderResourceAssociation: for attribute(s) ['resource_id'] with value(s) a186f98b-eb36-4baf-a753-e6178681eedb
2025-12-11 16:40:23.668 41 ERROR futurist.periodics
In my searches, this means that the Neutron service tries to insert data into the table but cannot because the data is already there.