Volume attach failed cause volume turn to available when delete the instance

[Wave Wang]

Description
===========
Using admin user to sign in. When a volume C is attached to instance A, if i try to attach it to another instance B, nova-api will
 raise an error "Invalid volume: Volume xxxx status must be available or downloading (HTTP 400) (Request-ID: xxxxx) (HTTP 400) (Request-ID: xxxxx)".At this time there is a deleted bdm record in nova for this failed attach.But then i delete instance B, the volume C (which is still attached to instance A) turned to available...making instance A can not use this volume.

Steps to reproduce
==================
* Using admin user to log in
* create an volume C
* attach volume C to instance A
* attach volume C to instance B, it failed and left a deleted bdm record in nova
* delete instance B, volume C alse become available and detached from instance A......

Expected result
===============
volume C should not become available and detached from instance A

Actual result
=============
volume C become available and detached from instance A

Environment
===========
rpm -qa|grep nova
openstack-nova-common-18.2.2-1.el7.noarch
openstack-nova-api-18.2.2-1.el7.noarch
python2-novaclient-11.0.0-1.el7.noarch
python-nova-18.2.2-1.el7.noarch

[Uggla (rene-ribaud)]

Thanks you for reporting this bug, I tried to reproduce using devstack and the Zed release.
But the bug can not be reproduced with this release on my side, unless if I'm wrong.

Please have a look at cmd log file attached and tell me if I miss something by setting back to new this bug.