Single lxd machine stuck on pending, Container started

This issue seems to be very active on jammy as well

We see this issue on juju 3.0rc2 as well: https://solutions.qa.canonical.com/testruns/testRun/285a9db6-3f91-4c23-926b-a0fc8c7ba389

@basdbruijne, the solqa results in #2 do not show the same bug as this.

In the case of #2, the container starts, as does the juju agent on the container, however it has errors.

$ grep pending juju-crashdump-openstack-2022-10-20-07.32.21/juju_status.txt
5/lxd/5 pending 10.246.167.51 juju-dd3b16-5-lxd-5 ubuntu:20.04 zone3 Container started

The machine agent shutdown down starting with?
2022-10-20 03:46:29 DEBUG juju.worker.dependency engine.go:616 "unconverted-api-workers" manifold worker stopped: agent should be terminated

Solutions QA is saying that they are also seeing failures to start containers on Jammy, and this is preventing them from doing their Openstack Jammy testing.

In testrun https://solutions.qa.canonical.com/testruns/testRun/fd79805c-8f0c-4965-af14-e01017439fe9 I looked around on the life env. Here, 2 machines are in this state:

```
Machine State Address Inst id Series AZ Message
0 started 10.246.167.190 solqa-lab1-server-07 jammy zone1 Deployed
0/lxd/0 started 10.246.167.149 juju-6ba804-0-lxd-0 jammy zone1 Container started
0/lxd/1 pending juju-6ba804-0-lxd-1 jammy zone1 Container started
0/lxd/2 started 10.246.164.253 juju-6ba804-0-lxd-2 jammy zone1 Container started
0/lxd/3 pending juju-6ba804-0-lxd-3 jammy zone1 Container started
0/lxd/4 started 10.246.166.148 juju-6ba804-0-lxd-4 jammy zone1 Container started
0/lxd/5 started 10.246.165.82 juju-6ba804-0-lxd-5 jammy zone1 Container started
0/lxd/6 started 10.246.167.96 juju-6ba804-0-lxd-6 jammy zone1 Container started
0/lxd/7 started 10.246.167.159 juju-6ba804-0-lxd-7 jammy zone1 Container started
0/lxd/8 started 10.246.166.215 juju-6ba804-0-lxd-8 jammy zone1 Container started
0/lxd/9 started 10.246.165.72 juju-6ba804-0-lxd-9 jammy zone1 Container started
0/lxd/10 started 10.246.164.203 juju-6ba804-0-lxd-10 jammy zone1 Container started
```

But logging on to the machines themselves shows no problem:
```
ubuntu@solqa-lab1-server-07:~$ sudo lxc list
To start your first container, try: lxc launch ubuntu:22.04
Or for a virtual machine: lxc launch ubuntu:22.04 --vm

+----------------------+---------+-----------------------+------+-----------+-----------+
| NAME | STATE | IPV4 | IPV6 | TYPE | SNAPSHOTS |
+----------------------+---------+-----------------------+------+-----------+-----------+
| juju-6ba804-0-lxd-0 | RUNNING | 10.246.173.8 (eth1) | | CONTAINER | 0 |
| | | 10.246.172.111 (eth1) | | | |
| | | 10.246.169.47 (eth0) | | | |
| | | 10.246.168.111 (eth0) | | | |
| | | 10.246.167.149 (eth2) | | | |
+----------------------+---------+-----------------------+------+-----------+-----------+
| juju-6ba804-0-lxd-1 | RUNNING | 10.246.176.28 (eth1) | | CONTAINER | 0 |
| | | 10.246.172.62 (eth0) | | | |
+----------------------+---------+-----------------------+------+-----------+-----------+
| juju-6ba804-0-lxd-2 | RUNNING | 10.246.172.251 (eth1) | | CONTAINER | 0 |
| | | 10.246.168.255 (eth0) | | | |
| | | 10.246.164.253 (eth2) | | | |
+----------------------+---------+-----------------------+------+-----------+-----------+
| juju-6ba804-0-lxd-3 | RUNNING | 10.246.169.48 (eth0) | | CONTAINER | 0 |
+----------------------+---------+-----------------------+------+-----------+-----------+
| juju-6ba804-0-lxd-4 | ...

Seen once more on a juju release run for version 2.9.38 here:

https://solutions.qa.canonical.com/v2/testruns/205c62cc-272b-4384-82c9-69a94db534a3/

Artifacts: https://oil-jenkins.canonical.com/artifacts/205c62cc-272b-4384-82c9-69a94db534a3/index.html

I confirm this happens too, consistently, when running the tutorial from https://juju.is/docs/sdk/build-and-deploy-minimal-machine-charm

$ juju version
2.9.38-ubuntu-amd64