However, NULL cell_mapping should not be an issue because for those instances who have an instance record in the the cell DB, but their instance_mapping has a NULL cell_mapping, meaning instances which are not yet scheduled, will have build_requests and as per (https://github.com/openstack/nova/blob/f902e0d5d87fb05207e4a7aca73d185775d43df2/nova/compute/api.py#L2402-L2408) the marker is set to None if its found in the BuildRequest table. So I don't fully understand what is the issue. @dberk, could you please attach the n-api logs if possible?
However, NULL cell_mapping should not be an issue because for those instances who have an instance record in the the cell DB, but their instance_mapping has a NULL cell_mapping, meaning instances which are not yet scheduled, will have build_requests and as per (https:/ /github. com/openstack/ nova/blob/ f902e0d5d87fb05 207e4a7aca73d18 5775d43df2/ nova/compute/ api.py# L2402-L2408) the marker is set to None if its found in the BuildRequest table. So I don't fully understand what is the issue. @dberk, could you please attach the n-api logs if possible?