Volume reserved after instance deleted

[Eric Xie]

Description
===========
I created one instance boot from volume.
After deleting the instance when it's building,
the volume's status was reserved.

Steps to reproduce
==================
* I created one instance with `openstack server create --volume vol-for-vm1-50G --flavor ecs_2C4G50G_general --network xtt-net-1 vm1`
* then I deleted the instance when it's status was build
| 3aa10ae1-2f3f-4c52-8f27-a557cf82de9e | vm1 | BUILD | None | NOSTATE | | | | ecs_2C4G50G_general | 5af94a8a-aab9-4ba5-bf52-ddb815218e61 | cn-north-3a | None | |
* then I showed the status of volume

Expected result
===============
The volume's status is available.

Actual result
=============
The volume's status is reserved.

Environment
===========
1. Exact version of OpenStack you are running. See the following
  list for all releases: http://docs.openstack.org/releases/
  # apt list --installed | grep nova

WARNING: apt does not have a stable CLI interface. Use with caution in scripts.

nova-api/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
nova-common/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
nova-conductor/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
nova-consoleauth/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
nova-consoleproxy/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
nova-doc/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
nova-placement-api/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
nova-scheduler/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed]
python-nova/xenial,xenial,now 2:17.0.7-6~u16.01 all [installed,automatic]
python-novaclient/xenial,xenial,now 2:9.1.1-1~u16.04+mcp6 all [installed,automatic]

# apt list --installed | grep cinder

WARNING: apt does not have a stable CLI interface. Use with caution in scripts.

cinder-backup/xenial,xenial,now 2:12.0.4-2~u16.04 all [installed]
cinder-common/xenial,xenial,now 2:12.0.4-2~u16.04 all [installed,automatic]
cinder-scheduler/xenial,xenial,now 2:12.0.4-2~u16.04 all [installed]
cinder-volume/xenial,xenial,now 2:12.0.4-2~u16.04 all [installed]
python-cinder/xenial,xenial,now 2:12.0.4-2~u16.04 all [installed,automatic]
python-cinderclient/xenial,xenial,now 1:3.5.0-1.0~u16.04+mcp5 all [installed,automatic]

[Matt Riedemann (mriedem)]

Do you see anything related in the logs like failing to unreserve the volume? If the server is not yet on a host the API will try to perform a "local delete" which should detach the volume:

https://github.com/openstack/nova/blob/17.0.7/nova/compute/api.py#L2085

Are you seeing warnings around that like it failed? Are there errors on the cinder side?

[Matt Riedemann (mriedem)]

Since this would be a race it's going to be hard to tell what happened without some investigation/logging to determine where it fails.

[Matt Riedemann (mriedem)]

I've marked this as Incomplete. If you can provide some logs with the volume ID for tracing then this could be debugged further.

[Launchpad Janitor (janitor)]

[Expired for OpenStack Compute (nova) because there has been no activity for 60 days.]