LXC termination fails using LVM for root

[Olivier]

Hello,

Terminating an instance running on LXC with LVM as instance root storage fails with an exception.

The LVM volume is wiped using dd, but it is not disconnected from qemu-nbd before the LVM volume gets removed.

The instance in itself on the compute node is stopped but still marked as Active, Deleting in nova. qemu-nbd is sill running.
After first failure, disconnecting the qemu-nbd process manually and relaunching the instance termination works

Log file in debug attached

Environment:
Ubuntu 12.04
OpenStack grizzly

Configuration for LVM:
libvirt_images_type=lvm
libvirt_images_volume_group=nova_local

Error: (The full stack can be find in attached log)
2013-04-25 12:02:59.119 10796 TRACE nova.openstack.common.rpc.amqp ProcessExecutionError: Unexpected error while running command.
2013-04-25 12:02:59.119 10796 TRACE nova.openstack.common.rpc.amqp Command: sudo nova-rootwrap /etc/nova/rootwrap.conf lvremove -f /dev/nova_local/instance-00000004_disk
2013-04-25 12:02:59.119 10796 TRACE nova.openstack.common.rpc.amqp Exit code: 5
2013-04-25 12:02:59.119 10796 TRACE nova.openstack.common.rpc.amqp Stdout: ''
2013-04-25 12:02:59.119 10796 TRACE nova.openstack.common.rpc.amqp Stderr: ' Can\'t remove open logical volume "instance-00000004_disk"\n'

[Alexey I. Froloff (raorn)]

Problem in nova/virt/libvirt/utils.py:find_disk(). It doesn't handle LVM volumes.