To Reproduce Steps to reproduce the behavior:

1. Check the containers of ovn-metadata haproxy running in a compute node.
2. Start a FFU from 16 to 17, run Compute node upgrade.
3. Check that any containers was created.
4. No error found.

Expected behavior

• Containers created and health or error log pointing the issue.

Device Info (please complete the following information):

• Hardware Specs: Total mem: 1547070 MB / Total CPU: 160
• OS Version: RHEL 9.2
• RHOSP: 17.1.4

Bug impact

• Customer is unable to use the updated computes.

Known workaround

• Downgrade to 17.1.3. Compute will work, but the agent will show as down in agent list.

Additional context

• Observed that some compute there were 2 images for openstack-neutron-metadata-agent-ovn. One at version 17.1.3 and another at version 17.1.4. It was tried to remove the image with 17.1.3 and restart the services tripleo_ovn_controller and tripleo_ovn_metadata_agent without success.