Description of problem: The revert operation of delete Cinder snapshot (The OnFailed), should not change the active flag of the disk. Version-Release number of selected component (if applicable): How reproducible: 100% Steps to Reproduce: 1. Create a VM with Cinder disk 2. Create a snapshot 3. Create another snapshot 4. delete the last snapshot and fail the operation of the delete Actual results: The first snapshot will become active and we will have two volumes which are active in the DB Expected results: We should not update the active volume Additional info:
Please consider that for this fix to work on multiple disks the patch with the following change-id I2ca19ec44cba371813381e8f8b076e463eda4173 should be backported. Until then, this fix should resolve the issue but mainly for one Cinder disk, for multiple disks there might be several corner cases which might not be supported.
Bug tickets must have version flags set prior to targeting them to a release. Please ask maintainer to set the correct version flags and only then set the target milestone.