fix: make sure binary install dir exists

[namsic]

I'm trying to install node_exporter on a remote host with a custom node_exporter_binary_install_dir.

- hosts: all
  roles:
    - prometheus.prometheus.node_exporter
  vars:
    node_exporter_binary_local_dir: /home/namsic/downloads/node_exporter-1.6.0.linux-amd64
    node_exporter_binary_install_dir: /home/namsic/tmp

But when I ran the above playbook, I got the following result:

TASK [prometheus.prometheus.node_exporter : Propagate locally distributed node_exporter binary] ****************************************************************************************************************************************************************************************************************************************************************************
fatal: [remote001]: FAILED! => {
    "changed": false,
    "checksum": "894db4cbc1bda8d686d73f33f22c57c861836c9d",
    "msg": "Destination directory /home/namsic/tmp does not exist"
}

If there are multiple target hosts, it would be nice to automatically create the directory if it does not exist.

If these changes are not appropriate, please let me know.

Thanks.

[namsic]

Can I just add the same change to every roles?

fix: ensure binary_install_dir before copy binary

Or should I do something else?

[gardar]

Can I just add the same change to every roles?

fix: ensure binary_install_dir before copy binary

Or should I do something else?

You can add a task to the install tasks file of the _common role which all the roles use.

Adding this task before the last task, the one named "Propagate binaries", in _common/tasks/install.yml would probably do the trick.

- name: "Make sure binary install dir exists"
  ansible.builtin.file:
    path: "{{ _common_binary_install_dir }}"
    mode: 0755
    owner: root
    group: root
  become: true
  tags:
    - "{{ ansible_parent_role_names | first | regex_replace(ansible_collection_name ~ '.', '') }}"
    - install
    - "{{ ansible_parent_role_names | first | regex_replace(ansible_collection_name ~ '.', '') }}_install"

[namsic]

Thanks for the detailed explanation. I changed it according to your comment.

[SuperQ]

I think we should use ansible.bultin.stat as a preflight check, rather than actually try and create the target directory.

[namsic]

Can someone help me?

I can't figure out how my ansible.builtin.stat changes relate to the error below.

INFO     Running alternative > destroy
INFO     Sanity checks: 'docker'
[WARNING]: Collection community.docker does not support Ansible version
2.11.12.post0

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

TASK [Destroy molecule instance(s)] ********************************************
fatal: [localhost]: FAILED! => {
   "msg": "Could not find imported module support code for ansible_collections.community.docker.plugins.modules.docker_container.
           Looked for (['ansible.module_utils.compat.version.LooseVersion', 'ansible.module_utils.compat.version'])"
}

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

WARNING  Retrying execution failure 2 of: ansible-playbook --inventory /root/.cache/molecule/alertmanager/alternative/inventory --skip-tags molecule-notest,notest /usr/local/lib/python3.6/dist-packages/molecule_docker/playbooks/destroy.yml
CRITICAL Ansible return code was 2, command was: ['ansible-playbook', '--inventory', '/root/.cache/molecule/alertmanager/alternative/inventory', '--skip-tags', 'molecule-notest,notest', '/usr/local/lib/python3.6/dist-packages/molecule_docker/playbooks/destroy.yml']

[gardar]

Can someone help me?

I can't figure out how my ansible.builtin.stat changes relate to the error below.

INFO     Running alternative > destroy
INFO     Sanity checks: 'docker'
[WARNING]: Collection community.docker does not support Ansible version
2.11.12.post0

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

TASK [Destroy molecule instance(s)] ********************************************
fatal: [localhost]: FAILED! => {
   "msg": "Could not find imported module support code for ansible_collections.community.docker.plugins.modules.docker_container.
           Looked for (['ansible.module_utils.compat.version.LooseVersion', 'ansible.module_utils.compat.version'])"
}

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

WARNING  Retrying execution failure 2 of: ansible-playbook --inventory /root/.cache/molecule/alertmanager/alternative/inventory --skip-tags molecule-notest,notest /usr/local/lib/python3.6/dist-packages/molecule_docker/playbooks/destroy.yml
CRITICAL Ansible return code was 2, command was: ['ansible-playbook', '--inventory', '/root/.cache/molecule/alertmanager/alternative/inventory', '--skip-tags', 'molecule-notest,notest', '/usr/local/lib/python3.6/dist-packages/molecule_docker/playbooks/destroy.yml']

That issue is unrelated to your change, the docker ansible collection which is used for the tests was updated recently and that broke compatibility with old ansible versions. I fixed the issue in #440

[github-actions[bot]]

Docs Build 📝

Thank you for contribution!✨

This PR has been merged and the docs are now incorporated into main: https://prometheus-community.github.io/ansible/branch/main