Quick question:

When LXC containers are created as part of the install process, what's the mechanism? Are they built from scratch? Are they downloaded? Are they cached on the ansible host or elsewhere and then moved to the target nodes for deployment? Or are they downloaded directly by the nodes?

I'm pretty sure the problem is somewhere here -- if the download fails silently, I could imagine the ansible scripts getting confused and thinking they have installed LXC images when they haven't, so the errors only show up when something tries to attach to the (nonexistent) container. If I run setup_openstack I get a similar failure, just on a different nonexistent LXC.

It's getting late here, I'm going to have another go at this tomorrow. Good night all! 

On Wed, Mar 12, 2025 at 9:19 PM Sarah Thompson <plodger@gmail.com> wrote:
The only thing I have in the variables config is the default:

---
# This file contains an example of the global variable overrides
# which may need to be set for a production environment.

## OpenStack public endpoint protocol
openstack_service_publicuri_proto: http
~                                          

Am investigating the LXC issues, not found anything yet, but I've only looked on the Ansible host so far. 

On Wed, Mar 12, 2025 at 9:11 PM Kerem Celiker <kmceliker@gmail.com> wrote:
++in addition, do u mind if you provide osp variable files on it to checking deep-dive about issue


Sevgiler,
Best,
Kerem ÇELİKER


On Thu, Mar 13, 2025 at 00:07 Sarah Thompson <plodger@gmail.com> wrote:
Hi Dmitriy,

Many thanks -- I've captured the complete output here:


Sarah

On Wed, Mar 12, 2025 at 8:59 PM Dmitriy Rabotyagov <noonedeadpunk@gmail.com> wrote:

Hey!

Sorry, can you please provide a little bit more output? As it's not even clear on what playbook/stage failure occurs.

As the error you have pasted can be a result of some prior failed task as well.

You can use https://paste.openstack.org/ to attach longer stdout


On Wed, 12 Mar 2025, 21:49 Sarah Thompson, <plodger@gmail.com> wrote:
Hi all,

I've been experimenting with openstack through a devstack instance, but I need to start looking at a more realistic deployment architecture. I have a 5 node Proxmox cluster available, so I've created 6 VMs, one to serve as a deployment host for Ansible etc, one as a controller node, and four compute nodes. These are all currently set up with Ubuntu 22 LTS and configured with 128GB of disk, 20GB of RAM and 12 cores. I've verified that Ansible works, LXC works correctly, etc.

I'm getting a failure when I'm running 

    openstack-ansible openstack.osa.setup_infrastructure

Specifically I get errors like:

PLAY [Gather repo facts] *******************************************************************************************************

TASK [Gathering Facts] *********************************************************************************************************
[WARNING]: Unhandled error in Python interpreter discovery for host openstack-controller-repo-container-e44299d5: unexpected
output from Python interpreter discovery
fatal: [openstack-controller-repo-container-e44299d5]: FAILED! =>
    ansible_facts: {}
    changed: false
    failed_modules:
        ansible.legacy.setup:
            ansible_facts:
                discovered_interpreter_python: /usr/bin/python3
            failed: true
            module_stderr: |-
                lxc-attach: openstack-controller-repo-container-e44299d5: attach.c: get_attach_context: 405 Connection refused - Failed to get init pid
                lxc-attach: openstack-controller-repo-container-e44299d5: attach.c: lxc_attach: 1469 Connection refused - Failed to get attach context
            module_stdout: ''
            msg: |-
                MODULE FAILURE
                See stdout/stderr for the exact error
            rc: 1
            warnings:
            - Platform unknown on host openstack-controller-repo-container-e44299d5 is
                using the discovered Python interpreter at /usr/bin/python3, but future
                installation of another Python interpreter could change the meaning of
                that path. See https://docs.ansible.com/ansible-core/2.17/reference_appendices/interpreter_discovery.html
                for more information.
    msg: |-
        The following modules failed to execute: ansible.legacy.setup


Inspecting the openstack-controller node manually with lxc-ls, there are no containers present. (I have manually created one and tested it successfully, so it's definitely not something weird with not being able to create/run containers).

Thank you in advance for your insights. :-)

Sarah
PS: When I have this working, next step is to put this on bare metal in a colo, assuming I can get it working.


--
[s]


--
[s]


--
[s]