> <@_oftc_eandersson:matrix.org> gboutry. Can you add this to the commit message of your PR?

 * I'm not sure this totally fixes the bug, it looks like this revision \[1\] is the real issue.

All tables created before this revision is applied have the tuple (charset, collation) = (utf8mb4, utf8mb4\_0900\_ai\_ci)

But the revision \[1\] sets `service_statuses` to (utf8mb3, utf8mb3\_general\_ci), and it is also applied as the new default of the database.

Then, this revision \[2\] (applied after the fix charset revision) is trying to create the `shared_zones` table with (utf8mb3, utf8mb3\_general\_ci), which fails due to table `zone` being (utf8mb4, utf8mb4\_0900\_ai\_ci)

I've tested to apply every migration while skipping [1] fix charset, and the db sync commands completes correctly, with all tables being my database default charset/collation: (utf8mb4, utf8mb4\_0900\_ai\_ci)

\[1\] (fix charset): https://opendev.org/openstack/designate/src/branch/stable/2023.1/designate/storage/impl\_sqlalchemy/alembic/versions/15b34ff3ecb8\_fix\_service\_charset.py
\[2\] (shared\_zones): https://opendev.org/openstack/designate/src/branch/stable/2023.1/designate/storage/impl\_sqlalchemy/alembic/versions/b20189fd288e\_shared\_zone.py