Description of problem: $subject, we've seen a number of examples recently where users have called reset-state [1][2][3] without updating the Nova DB and leaving the volume still attached there. This is an admin only command/API that really should only be used in conjunction with Nova DB updates (I am thinking about introducing force detach API FWIW) when users understand the issue and the required steps to fix it. A clear warning being printed should help avoid users getting stuck in weird and wonderful corners going forward. [1] https://bugzilla.redhat.com/show_bug.cgi?id=1906941 [2] https://bugzilla.redhat.com/show_bug.cgi?id=1887383 [3] https://bugzilla.redhat.com/show_bug.cgi?id=1883583 Version-Release number of selected component (if applicable): OSP 16.1 How reproducible: Always Steps to Reproduce: 1. Attempt to call `cinder reset-state` Actual results: Called without printing a clear warning etc. Expected results: Print a clear warning etc. Additional info:
This upstream spec is aiming to resolve issues like this: https://review.opendev.org/c/openstack/cinder-specs/+/682456
Should we consider this a bugfix or a feature? The spec above is not approved for Wallaby right now, and depending on the fix it may be complicated to backport the changes.
The WIP upstream spec is now https://review.opendev.org/c/openstack/cinder-specs/+/766716 - this is going to be a feature.