The following tempest tests are failing:
tempest.api.volume.test_volume_delete_cascade.VolumesDeleteCascade.test_volume_from_snapshot_cascade_delete
tempest.api.volume.test_volumes_snapshots.VolumesSnapshotTestJSON.test_volume_from_snapshot
tempest.scenario.test_volume_boot_pattern.TestVolumeBootPattern.test_create_ebs_image_and_check_boot
tempest.scenario.test_volume_boot_pattern.TestVolumeBootPattern.test_create_server_from_volume_snapshot
tempest.scenario.test_volume_boot_pattern.TestVolumeBootPattern.test_volume_boot_pattern
All these tests fails because cinder isn't being able to attach a volume to a vm from snapshot.
Steps for these tests are:
* Create in Cinder some bootable volume importing a Glance image - Ok
* Boot an instance from the bootable volume - Ok
* Write content to the volume - Ok
* Delete an instance and Boot a new instance from the volume - Ok
* Check written content in the instance - Ok
* Create a volume snapshot while the instance is running - Ok
* Boot an additional instance from the new snapshot based volume - FAIL
* Check written content in the instance booted from snapshot - Not executed
Here's the log from /var/log/cinder/volume.log
http://paste.openstack.org/raw/614791/
This was caught today in the tripleo periodic master job running tempest. The error don't appear in the ocata periodic job.
Adding cinder, in the recent week(s) there were multiple issues with thin-lvm reverts/new fixes , so the root cause of this issue is likely in cider.