Resize operation cannot be performed by admin user on tenant cluster

[Andrew Bonney]

Assuming this isn't by design, I've found that issuing the following command as an admin user against a tenant cluster does not work:

openstack coe cluster resize <cluster_name> 3

This results in an error such as:

Cluster fdf0a32a-164a-411b-8fb0-150281162b0c could not be found (HTTP 404) (Request-ID: req-512ab042-81f4-4598-9514-49213b9790d2)

Having traced this through to the Magnum codebase, the call to resize includes https://github.com/openstack/magnum/blob/master/magnum/db/sqlalchemy/api.py#L233C26-L233C26, which adds tenant filters (https://github.com/openstack/magnum/blob/master/magnum/db/sqlalchemy/api.py#L129C52-L129C52), but 'context.all_tenants' is set to False, so the request is attempted against the admin user's project and results in the 404 above.

Direct command line calls to list and show clusters work correctly, so I'm not sure why the context is set differently in this case. Unfortunately as there isn't an '--all-tenants' argument available (as there is for 'openstack coe quotas list' for example), it doesn't appear to be possible to work around this.

I haven't been able to test cluster upgrades yet, so I can't say if this is unique to resize operations of if it might apply to other commands as well.