Open Stack

Hi,
while trying to reproduce the issue mentioned in [1] I stumbled upon  
[2], I commented there but because of the "wishlist" importance I  
don't expect too much attention.
I have a production cloud with Victoria where I see the same issue as  
described in [2], trying to retype an in-use volume results in a stack  
trace in cinder:

2023-06-20 12:45:52.900 1085571 ERROR oslo_messaging.rpc.server  
keystoneauth1.exceptions.http.BadRequest: Expecting to find domain in  
project. The server could not comply with the request since it is  
either malformed or otherwise incorrect. The client is assumed to be  
in error. (HTTP 400) (Request-ID:  
req-6b412e96-a735-4b94-a5e6-8252dd949297)

I have also a test cloud installed the same way as the production  
cloud with V and upgraded to Wallaby where this stack trace does not  
occur, retyping an in-use volume just works there. I tried the  
described workaround in [1] and added a [nova] section to the  
cinder.conf, and that worked. There is no such section in the W cloud.  
I checked the Wallaby release notes and the only thing related to  
retype with ceph as backend is [3]. Could that make the difference here?
One difference between V and W I noticed was that in V I see two  
volumes with the same name (but a different UUID, of course) during  
the retype process while in W I see only one volume at all times. So I  
assume that W handles the retype differently and it doesn't require a  
nova section.
I'd appreciate any comment.

Thanks,
Eugen

[1]  
https://lists.openstack.org/pipermail/openstack-discuss/2023-June/034160.html
[2] https://bugs.launchpad.net/cinder/+bug/1941062
[3] https://bugs.launchpad.net/cinder/+bug/1886543