Comment 2 for bug 1682634

First of all, I have to say that attached commands output doesn't indicate any issues: volume IDs are stored in volume_id row of volume_attachment table, not id row.

On the other hand, I was unable to reproduce the issue by retyping attached volume in our lab (I hadn't changed volume driver, so no migration was needed). As a result, I have to say that steps to reproduce are not correct.

Scott, can you please improve bug description and add the following information:

- Describe the cloud architecture, specify used storages and their cinder drivers
- Describe volume types used
- Enable nova and cinder debug, collect logs during retype
- Use openstack commands (not DB requests) to get information.

From my point of view, current bugs status is Invalid, since it is impossible to reproduce it with described steps and attached command output is 100% correct (doesn't describe any bug).