Issue with ansible to access lxd container

LXD Support
1

I have ansible on the same lxd server, the purpose is I want ansible to manage containers
so I created container
and a dynamic inventory (so that I don have to type the name of container each time)
so here is the dynamic inventory

plugin: community.general.lxd
url: unix:/var/snap/lxd/common/lxd/unix.socket
state: RUNNING

here is the output of the inventory

ansible-inventory -i lxd.yaml --graph
@all:
  |--@ungrouped:
  |  |--nginx-1

I ran ansible -i lxd.yaml -m ping all
and got this output

nginx-1 | UNREACHABLE! => {
    "changed": false,
    "msg": "Failed to connect to the host via ssh: root@container-ip: Permission denied (publickey).",
    "unreachable": true
}

then I have this inventory

all :
  vars : 
    ansible_connection : lxd 
    ansible_user : root 
    ansible_become : no 
  children : 
    local : 
      vars : 
        ansible_lxd_remote : local 
        ansible_lxd_project : default
      hosts :
        nginx-1

and

ansible -i hosts.yaml -m ping all
nginx-1 | SUCCESS => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/bin/python3"
    },
    "changed": false,
    "ping": "pong"
}

so my guess is in the dynamic inventory use ssh to access the container but failed on the second part I specify to use ansible_connection=lxd so ansible was able to access the container
is there any way to make ansible use lxd connection on the dynamic inventory

2

There seems to be an issue with ansible-inventory as it reports the ansible_connection as ssh for all running instances (and local for those stopped). You can see it with ansible-inventory -i lxd.yml --list:

{
    "_meta": {
        "hostvars": {
            "c1": {
                "ansible_connection": "ssh",
                "ansible_host": "172.24.26.109",
                "ansible_lxd_os": "ubuntu",
                "ansible_lxd_profile": [
                    "default"
                ],
                "ansible_lxd_project": "default",
                "ansible_lxd_release": "noble",
                "ansible_lxd_state": "running",
                "ansible_lxd_type": "container"
            },
            "j1": {
                "ansible_connection": "local",
                "ansible_lxd_os": "ubuntu",
                "ansible_lxd_profile": [
                    "default"
                ],
                "ansible_lxd_project": "default",
                "ansible_lxd_release": "noble",
                "ansible_lxd_state": "stopped",
                "ansible_lxd_type": "container"
            },
        }
    },
    "all": {
        "children": [
            "ungrouped"
        ]
    },
    "ungrouped": {
        "hosts": [
            "j1",
            "c1",
        ]
    }
}

FYI, my lxd.yml file:

# Usage: ansible-inventory -i lxd.yml --list
plugin: community.general.lxd
url: unix:/var/snap/lxd/common/lxd/unix.socket

That said, if you are OK with a static inventory, it works well:

ansible.cfg:

$ cat ansible.cfg
[defaults]
inventory = hosts.yml
timeout = 30
forks = 10
# avoids creating multiple sessions with systemd-logind
pipelining = True

[inventory]
enable_plugins = yaml, community.general.lxd

hosts.yml:

---
all:
  vars:
    ansible_connection: lxd
    ansible_lxd_project: default
    ansible_lxd_remote: local
    ansible_user: root
    ansible_become: false
  hosts:
    c1:
    j1:

books/ping.yml:

---
- name: ping test
  hosts: all
  tasks:
    - name: test connectivity
      ansible.builtin.ping:

$ ansible-playbook books/ping.yml 

PLAY [ping test] *******************************************************************************************************************************************************************************************************************************************************************

TASK [Gathering Facts] *************************************************************************************************************************************************************************************************************************************************************
[WARNING]: Unhandled error in Python interpreter discovery for host j1: instance not running: j1
fatal: [j1]: UNREACHABLE! => {"changed": false, "msg": "instance not running: j1", "unreachable": true}
ok: [c1]

TASK [test connectivity] ***********************************************************************************************************************************************************************************************************************************************************
ok: [c1]

PLAY RECAP *************************************************************************************************************************************************************************************************************************************************************************
c1                         : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
j1                         : ok=0    changed=0    unreachable=1    failed=0    skipped=0    rescued=0    ignored=0