Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Help needed on configuration #17

Open
bitchecker opened this issue Aug 7, 2024 · 29 comments
Open

Help needed on configuration #17

bitchecker opened this issue Aug 7, 2024 · 29 comments
Labels
question Further information is requested

Comments

@bitchecker
Copy link

Hi,
there is some way to debug what is happening when molecule test command is executed with this driver?

I'm using a simple configuration:

driver:
  name: molecule-proxmox
  options:
     api_host: xxxxxxx
     api_user: root@pam
     api_password: "********"
     node: pve
     ssh_user: root
     ssh_identity_file: /path/to/id_rsa
platforms:
  - name: test01
    template_name: template01
    ciuser: root
    cipassword: root
    ipconfig:
      ipconfig0: 'ip=x.y.z.k/24,gw=x.y.z.1'
    nameservers:
      - 1.1.1.1

but when I try to exec a simple test (using default scenario auto-generated files) nothing happens:

WARNING  Driver molecule-proxmox does not provide a schema.
INFO     default scenario test matrix: dependency, cleanup, destroy, syntax, create, prepare, converge, idempotence, side_effect, verify, cleanup, destroy
INFO     Performing prerun with role_name_check=0...
INFO     Running default > dependency
WARNING  Skipping, missing the requirements file.
WARNING  Skipping, missing the requirements file.
INFO     Running default > cleanup
WARNING  Skipping, cleanup playbook not configured.
INFO     Running default > destroy

PLAY [Destroy] *****************************************************************

TASK [Populate instance config] ************************************************
ok: [localhost]

TASK [Dump instance config] ****************************************************
skipping: [localhost]

PLAY RECAP *********************************************************************
localhost                  : ok=1    changed=0    unreachable=0    failed=0    skipped=1    rescued=0    ignored=0

INFO     Running default > syntax

playbook: molecule/default/converge.yml
INFO     Running default > create

PLAY [Create] ******************************************************************

TASK [Populate instance config dict] *******************************************
skipping: [localhost]

TASK [Convert instance config dict to a list] **********************************
skipping: [localhost]

TASK [Dump instance config] ****************************************************
skipping: [localhost]

PLAY RECAP *********************************************************************
localhost                  : ok=0    changed=0    unreachable=0    failed=0    skipped=3    rescued=0    ignored=0

INFO     Running default > prepare

PLAY [Prepare] *****************************************************************

TASK [Waiting for instance ssh connection.] ************************************

No connections on proxmox server, so no machines are created...
@IamLunchbox
Copy link
Collaborator

IamLunchbox commented Aug 11, 2024 •
edited
Loading

Are you sure that you are able to reach the spun up machine on the default ssh port? It seems like ansible is not able to reach the machine via ssh on port 22. In your other issue you mentioned using an alternative port. Could this be an issue here?

I would furthermore suggest you run molecule with more verbosity (-v) to find out what molecule is trying to do there. Maybe molecule uses an (for you at least) unexpected ansible.cfg file and the therein used ssh-port to connect.

You can actually override the connection port (and other parameters) in your molecule config:

driver:
  name: molecule-proxmox
  options:
  [...]
platforms:
  [...]
provisioner:
  name: ansible
  config_options:
    ssh-connection:
      host_key_checking: false

@bitchecker
Copy link
Author

bitchecker commented Aug 11, 2024 •
edited
Loading

Hi, thanks for reply!

Nope, is not related to other raised issue, after raising that, I updated connection configs and reach directly the pve server using standard 22 and 8006 ports.

I'll try with -v option asap.

@bitchecker
Copy link
Author

Hi @IamLunchbox,
I tried with more verbosity and I can see that molecule is waiting to connect to the new VM but no VM are going to start so, I'm pretty sure that something should be configured in create.yml and destroy.yml files, but I can find any documentation on how to compile them with this provider.

@IamLunchbox
Copy link
Collaborator

IamLunchbox commented Aug 12, 2024 via email

@bitchecker
Copy link
Author

The big problem is that pve is not starting a new vm from the template...so it seems that nothing happens.
I'm asking for prepare.yml and destroy.yml because of that.

@IamLunchbox
Copy link
Collaborator

IamLunchbox commented Aug 12, 2024 via email

@bitchecker
Copy link
Author

molecule.yml file:

---
driver:
  name: molecule-proxmox
  options:
    api_host: xxxxxxxxx
    api_user: root@pam
    api_password: xxxxxxx
    node: proxmox
    ssh_user: root
    ssh_identity_file: /path/to/ssh/private/key
platforms:
  - name: molecule
    template_name: Alma9-template
    ciuser: root
    cipassword: root
    ipconfig:
      ipconfig0: 'ip=192.168.1.244/24,gw=192.168.1.1'
    nameservers:
      - 1.1.1.1
provisioner:
  name: ansible
  config_options:
    ssh-connection:
      host_key_checking: false
      ansible_ssh_common_args: '-p ssh-port'
  lint:
    name: ansible-lint

molecule -v test output:

PLAY [Create] ******************************************************************

TASK [Populate instance config dict] *******************************************
skipping: [localhost] => {"changed": false, "false_condition": "server.changed | default(false) | bool", "skip_reason": "Conditional result was False"}

TASK [Convert instance config dict to a list] **********************************
skipping: [localhost] => {"changed": false, "false_condition": "server.changed | default(false) | bool", "skip_reason": "Conditional result was False"}

TASK [Dump instance config] ****************************************************
skipping: [localhost] => {"changed": false, "false_condition": "server.changed | default(false) | bool", "skip_reason": "Conditional result was False"}

PLAY RECAP *********************************************************************
localhost                  : ok=0    changed=0    unreachable=0    failed=0    skipped=3    rescued=0    ignored=0

INFO     Running default > prepare
Using /molecule/default/ansible.cfg as config file

PLAY [Prepare] *****************************************************************

TASK [Waiting for instance ssh connection.] ************************************

As suggested, all files are at default status, only molecule.yml is updated with connection parameters.

@bitchecker
Copy link
Author

bitchecker commented Aug 12, 2024 •
edited
Loading

Playing with create.yml and using proxmox_kvm module is it possible to manage the VM creation, also if after that I need to understand how to add them to molecule inventory.

@meffie
Copy link
Owner

meffie commented Aug 13, 2024

Thanks bitchecker. What did you need to change in create.yml to fix your issue?

@bitchecker
Copy link
Author

Hi @meffie,
in create.yml I'm adding something like that to manage the VMs creation:

- name: Clone cloud-init template
  community.general.proxmox_kvm:
    api_user: root@pam
    api_password: xxxxxxxxx
    api_host: xxxxxxxxxx
    node: proxmox
    vmid: <template-id>
    newid: <new-id> # without other steps fails, some proxmox_kvm module bug?
    clone: Alma9-template
    name: molecule
    storage: local-zfs
    timeout: 90
    net:
      net0: 'virtio,bridge=vmbr0'

- name: Update VM configuration
  community.general.proxmox_kvm:
    api_user: root@pam
    api_password: xxxxxxxxx
    api_host: xxxxxxxxxx
    node: proxmox
    vmid: <new-id> # without other steps fails, some proxmox_kvm module bug?
    cores: 2
    memory: 4096
    ide:
      ide2: 'local-zfs:cloudinit,media=cdrom' # this is not working!
    ciuser: root
    cipassword: root
    nameservers: 1.1.1.1
    ipconfig:
      ipconfig0: 'ip=<ip-address>/24,gw=<gateway-address>'
    tags: molecule
    update: true

- name: Start VM
  community.general.proxmox_kvm:
    api_user: root@pam
    api_password: xxxxxxxxx
    api_host: xxxxxxxxxx
    node: proxmox
    vmid: <new-id> # without other steps fails, some proxmox_kvm module bug?
    state: started

What is missing is that at the end of these steps, the vm should add into the "molecule" inventory, that is this (default code):

  - name: Create instance config
    when: server.changed | default(false) | bool  # noqa no-handler
    block:
      - name: Populate instance config dict  # noqa jinja
        ansible.builtin.set_fact:
          instance_conf_dict: {}
          # instance': "{{ }}",
          # address': "{{ }}",
          # user': "{{ }}",
          # port': "{{ }}",
          # 'identity_file': "{{ }}", }
        with_items: "{{ server.results }}"
        register: instance_config_dict

      - name: Convert instance config dict to a list
        ansible.builtin.set_fact:
          instance_conf: "{{ instance_config_dict.results | map(attribute='ansible_facts.instance_conf_dict') | list }}"

      - name: Dump instance config
        ansible.builtin.copy:
          content: |
            # Molecule managed

            {{ instance_conf | to_json | from_json | to_yaml }}
          dest: "{{ molecule_instance_config }}"
          mode: "0600"

And I'm still trying to understand how to attach new code with default one.

In destroy.yml instead, I added this:

- name: Stop VM
  community.general.proxmox_kvm:
    api_user: root@pam
    api_password: xxxxxxxx
    api_host: xxxxxxx
    node: proxmox
    vmid: 9000
    state: absent
    force: true

As I said in the "code comments" the Cloud-Init directives are failing and no logs are provided.

@bitchecker
Copy link
Author

@meffie , @IamLunchbox any news on this?

@IamLunchbox
Copy link
Collaborator

IamLunchbox commented Aug 24, 2024 via email

@bitchecker
Copy link
Author

for molecule:

molecule==24.8.0
molecule-proxmox==1.0.0

for galaxy community.general should not be necessary because everything is already available, but:

community.general:9.3.0

proxmox: 8.2.4

@IamLunchbox
Copy link
Collaborator

I tried to reproduce your problem with a clean virtualenv, your specific versions and only the packages molecule-proxmox, molecule, ansible, requests. I was not able to reproduce your issue.

I suspect the main issue may be, that molecule does not seem to pick up the molecule-proxmox create.yml-playbook. The conditional check when: server.changed is not reached my execution.

But there was a similar issue in my testrun, which could be related. In one execution, Molecule somehow did not pickup the correct path for the molecule-proxmox library and did try to use and old locally installed molecule-version. This was only fixed after closing the shell and reactivating the virtualenv afterwards. But then it was working reliably.

My successful execution logged:
1 plays in /home/ubuntu/Repos/testrepo/venv/lib/python3.10/site-packages/molecule_proxmox/playbooks/create.yml.

My failed execution logged:
1 plays in /home/ubuntu/.local/lib/python3.10/site-packages/molecule_proxmox/playbooks/create.yml - and it failed because that version did not have the pool-parameter yet (molecule-proxmox 0.9).

If you not did so before, check with more verbosity (-vv) what paths molecule identified for the given create and destroy playbook and use a clean virtualenv for your tests.

Lastly, maybe try to remove all extra config items from your molecule config.Just to make sure a misplaced dict item is not throwing something off - even though I don't expect this to be the case. When you use the maximum verbosity (-vvvv) you'll see the config items molecule created and used, e.g. to pass them on to proxmox_kvm - maybe this will give you a hint on whats going on.

@bitchecker
Copy link
Author

Hi,
for what I can see, the create.yml is completely skipped (with default file):

PLAYBOOK: create.yml ***********************************************************
Positional arguments: /home/bitchecker/molecule/molecule/default/create.yml
verbosity: 4
connection: ssh
become_method: sudo
tags: ('all',)
skip_tags: ('notest', 'molecule-notest')
inventory: ('/dev/shm/bitchecker/molecule/molecule/default/inventory',)
forks: 50

PLAY [Create] ******************************************************************

TASK [Populate instance config dict] *******************************************
task path: /home/bitchecker/molecule/molecule/default/create.yml:13
skipping: [localhost] => {
    "changed": false,
    "false_condition": "server.changed | default(false) | bool",
    "skip_reason": "Conditional result was False"
}

TASK [Convert instance config dict to a list] **********************************
task path: /home/bitchecker/molecule/molecule/default/create.yml:24
skipping: [localhost] => {
    "changed": false,
    "false_condition": "server.changed | default(false) | bool",
    "skip_reason": "Conditional result was False"
}

TASK [Dump instance config] ****************************************************
task path: /home/bitchecker/molecule/molecule/default/create.yml:28
skipping: [localhost] => {
    "changed": false,
    "false_condition": "server.changed | default(false) | bool",
    "skip_reason": "Conditional result was False"
}

PLAY RECAP *********************************************************************
localhost                  : ok=0    changed=0    unreachable=0    failed=0    skipped=3    rescued=0    ignored=0

So, the main issue starts from here.

For converge.yml he tries to connect to the name that you give to the platform in molecule.yml so I can see in my logs:

sending connection check: [b'ssh', b'-vvvv', b'-C', b'-o', b'ControlMaster=auto', b'-o', b'ControlPersist=60s', b'-o', b'StrictHostKeyChecking=no', b'-o', b'KbdInteractiveAuthentication=no', b'-o', b'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', b'-o', b'PasswordAuthentication=no', b'-o', b'ConnectTimeout=10', b'-o', b'ControlPath="/home/bitchecker/.ansible/cp/%h-%p-%r"', b'-O', b'check', b'molecule']

so, if you're not in the same network of the proxmox server, and you haven't a DNS that can solve that name, you'll never reach that guest (that will never be reached because if you don't update the create.yml file, it will never be created)

@IamLunchbox
Copy link
Collaborator

IamLunchbox commented Aug 26, 2024 via email

@bitchecker
Copy link
Author

When I created the env, I created a new scenario, and I'm using the default create.yml and others. The only files that I changed were converge.yml and molecule.yml.

In on previous test, I also completely deleted all files and kept only those two.

I still think that if you're not on the same PVE network that uses default ports and with guests on the bridge in the same network, all these stuff will never work...so also if your PVE is hosted (with natted guests) you'll be never able to use this.

@IamLunchbox
Copy link
Collaborator

IamLunchbox commented Aug 26, 2024 •
edited
Loading

In on previous test, I also completely deleted all files and kept only those two.

Good to know! I will test if molecule-docker works with the default files generated by molecule. If it does, this will probably stop this project being ever merged into the molecule-plugins repo and we should check, what keeps the default config from being run with molecule-proxmox.

I still think that if you're not on the same PVE network that uses default ports and with guests on the bridge in the same network, all these stuff will never work...

This is incorrect. You don't need to use the same network, if you use this plugin. I always work through VPN without hostname resolution. Molecule saves the ip associated IP-addresses, not only the hostname, into its cache file:

ubuntu@adm: cat /home/ubuntu/.cache/molecule/test/default/inventory/ansible_inventory.yml         
# Molecule managed

---
all:
  hosts:
    test: &id001
      ansible_host: 10.0.0.233
      ansible_port: 22
      ansible_private_key_file: redacted
      ansible_ssh_common_args: -o UserKnownHostsFile=/dev/null -o ControlMaster=auto
        -o ControlPersist=60s -o ForwardX11=no -o LogLevel=ERROR -o IdentitiesOnly=yes
        -o StrictHostKeyChecking=no
      ansible_user: test
      connection: ssh
  vars:
    molecule_ephemeral_directory: '{{ lookup(''env'', ''MOLECULE_EPHEMERAL_DIRECTORY'')
      }}'
    molecule_file: '{{ lookup(''env'', ''MOLECULE_FILE'') }}'
    molecule_instance_config: '{{ lookup(''env'', ''MOLECULE_INSTANCE_CONFIG'') }}'
    molecule_no_log: '{{ lookup(''env'', ''MOLECULE_NO_LOG'') or not molecule_yml.provisioner.log|default(False)
      | bool }}'
    molecule_scenario_directory: '{{ lookup(''env'', ''MOLECULE_SCENARIO_DIRECTORY'')
      }}'
    molecule_yml: '{{ lookup(''file'', molecule_file) | from_yaml }}'
ungrouped:
  hosts:
    test: *id001
  vars: {}

so also if your PVE is hosted (with natted guests) you'll be never able to use this.

I think you can safely assume that nobody tested the case you outlined yet. But if you think using a different port keeps you from doing so with molecule-proxmox 1.0: This is not the case. Currently, use ansible_port to use the given port for ALL platforms of that scenario.

name: ansible
  connection_options:
    ansible_port: 23

And because you said never: If you want to have an alternative port per platform, please create a PR. I am sure michael is open for improvements, especially when the change is within the spec of the proxmox_kvm module.

@IamLunchbox
Copy link
Collaborator

I rechecked the behaviour with molecule init scenario test and the podman driver. It works if I just provide a molecule.yml and a converge.yml.

After initiating the scenario, which leads to three default files being created (create, prepare, destroy) the same behavior occurs as described by you.

Therefore, I assume that the skipping of instance creation when using the default create.yml does not seem to be a molecule-proxmox specific error.

@IamLunchbox
Copy link
Collaborator

@bitchecker could you fix this problem and/or is it ok to close this issue?

@bitchecker
Copy link
Author

Hi,
according to my point of view avoiding configuring the create file should be the default as the tool should directly take care of creating guests on proxmox in order to run tests on the Ansible code, as I showed in the previous comments, it does not do this by default unless major changes are made by the user.

If this is the expected behavior, you can proceed to close the issue.

@meffie
Copy link
Owner

meffie commented Sep 20, 2024

Sorry for the delay. I was traveling when you replied back in August and missed that reply. I'll take a closer look at the info you provided.

@IamLunchbox
Copy link
Collaborator

When I created the env, I created a new scenario, and I'm using the default create.yml and others. The only files that I changed were converge.yml and molecule.yml.

@bitchecker can you confirm your issue, if you only provide converge.yml and molecule.yml in your scenario? Please delete all other files from the scenario directory.

@bitchecker
Copy link
Author

Hi,
yes, I'm using only converge and molecule but I would like to clarify a missing point:
Is this software designed to test Ansible code interacting with proxmox and automatically provision guests, or I'm wrong?

If I'm right, this issue is just reporting that guests are not provisioned, and I provided all the requested logs. If I'm going to delete create and destroy, I'm telling that creation and deletion of guests are not managed by this tool... Nothing else.

@IamLunchbox
Copy link
Collaborator

IamLunchbox commented Sep 22, 2024 via email

@bitchecker
Copy link
Author

So no, this code is not designed to make roles interact with proxmox, but rather with ephemeral vms.

Sorry, when I said this?

Molecule tests ansible roles. To do so, it needs a system to configure. The system is provided by a driver (docker, azure, podman) and each single system is a plattform.
Molecule proxmox is a driver, which provides you an environment to test a role. To do so, it reaches out to proxmox, clones a vm and connects to it via ssh.

It's exactly what I've said in my previous comment:

Is this software designed to test Ansible code interacting with proxmox and automatically provision guests, or I'm wrong?

And you're just confirming that this driver should provide the environment, provisioning guests on proxmox. I'm telling that this not happens.

@IamLunchbox
Copy link
Collaborator

Is this software designed to test Ansible code interacting with proxmox and [...]

It may be a translation error, but this is what I was referring to.

Anyways, all the logs you provided above (1 and 2) reference the default create step from molecule scenario init - not the create step from molecule-proxmox.

But I don't have an idea, why these steps happen in your instance.

@IamLunchbox IamLunchbox added the question Further information is requested label Oct 3, 2024
@IamLunchbox
Copy link
Collaborator

I'll close this due to inactivity.

@meffie
Copy link
Owner

meffie commented Oct 26, 2024

Sorry, I've been inactive on this one. Looks like there is more to do here.

@meffie meffie reopened this Oct 26, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
question Further information is requested
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@meffie @bitchecker @IamLunchbox