lxd doesn't provision instances on first bootstrap in new xenial image

I have hit this every time I bootstrap the lxd provider on a new xenial instance. The bootstrap works fine, but when I try to deploy anything into either the default or admin model, a new instance is never provisioned, and the unit just hangs in "Waiting for agent initialization to finish":

ubuntu@ip-172-31-18-222:~$ juju2 switch lxd:default
local.lxd:admin -> local.lxd:default
ubuntu@ip-172-31-18-222:~$ juju2 status
[Services]
NAME STATUS EXPOSED CHARM
ubuntu unknown false cs:trusty/ubuntu-7

[Units]
ID WORKLOAD-STATUS JUJU-STATUS VERSION MACHINE PORTS PUBLIC-ADDRESS MESSAGE
ubuntu/0 unknown allocating Waiting for agent initialization to finish

[Machines]
ID STATE DNS INS-ID SERIES AZ

If I go into the database, the only machine in the machines collection is the state server.

juju:PRIMARY> db.machines.find().count()
1

I cannot destroy / kill the controller because it loops forever waiting for the service to be destroyed:
ubuntu@ip-172-31-18-222:~$ juju2 destroy-controller lxd --destroy-all-models
WARNING! This command will destroy the "local.lxd" controller.
This includes all machines, services, data and other resources.

Continue [y/N]? y
Destroying controller "local.lxd"
Waiting for hosted model resources to be reclaimed.
Waiting on 1 model, 2 services
Waiting on 1 model, 2 services
Waiting on 1 model, 2 services
Waiting on 1 model, 2 services
<repeated until I kill it>

Oddly enough, if I bootstrap a second lxd controller, I can deploy just fine.