Customer team, with help from our consulting team, tried to perform FFU 16.2 to 17.1 on their sandbox environment in order to validate it for production environment. They have a multi cell environment, so they have 3 stacks namely overcloud, cell1, and cell2.
They are using the following documentation for the FFU.

https://docs.redhat.com/en/documentation/red_hat_openstack_platform/17.1/html-single/framework_for_upgrades_16.2_to_17.1/index#upgrading-rhosp-on-all-nodes-in-each-stack_overcloud-upgrade

During cell1 upgrade in multi cell environnement (8.1 in the documentation), ansible tried to run `podman exec nova_conductor nova-manage db online_data_migration` through cell1_controller with a delegate_host, and it fails with the error [1].
They already tried to apply this solution but same issue:

https://access.redhat.com/solutions/7106823

So, the consulting team working on the customer project deleted (dropped) the database `nova_cell0` and then tried the online_data_migration, but got the same issue.

Then they deleted the cell0 cell through nova-manage cli commands [2], after that they got no more issue relaunching online_data_migration command.

So, is it a bug at the tripleo end?

Also, could you shed more light on whether or not the aforementioned workaround can be used? Will the above workaround cause any behavioral difference among the Nova or other components?

Regards

[1]

```

ERROR nova.cmd.manage sqlalchemy.exc.ProgrammingError: (pymysql.err.ProgrammingError) (1146, "Table 'nova_cell0.virtual_interfaces' doesn't exist")

```

[2]

```

sudo podman exec nova_conductor nova-manage cell_v2 list_cells
sudo podman exec nova_api nova-manage cell_v2 delete_cell --cell_uuid 00000000-0000-0000-0000-000000000000

```
o REPRODUCER STEPS:

Follow the documentation until the section: https://docs.redhat.com/en/documentation/red_hat_openstack_platform/17.1/html-single/framework_for_upgrades_16.2_to_17.1/index#upgrading-rhosp-on-all-nodes-in-each-stack_overcloud-upgrade