如何在CloudStack的cs_instance模块中按变量定义状态添加/移除标签

I'm using the Ansible cs_instance module to create CloudStack instances, here's my sample code:

- name: create or update instance
  cs_instance:
    display_name: web-vm-1
    template: Linux Debian 7 64-bit
    service_offering: Tiny
    ssh_key: john@example.com
    tags:
      - { key: admin, value: john }
      - { key: group, value: "{{ vm_group_tab }}" }
    user_data: |
      #cloud-config
      packages:
        - nginx

I want to add or remove the group tag depending on whether the vm_group_tab variable is defined. Specifically, when vm_group_tab isn't defined, I only want the admin tag to be present. How can I achieve this?

Great question! There are a couple of clean, straightforward ways to handle this dynamic tag list in Ansible, depending on your preference:

Method 1: Jinja2 Conditionals Directly in the Task #

The simplest approach is to use Jinja2's conditional syntax to include the group tag only when vm_group_tab is defined. This keeps everything contained within your cs_instance task:

- name: create or update instance
  cs_instance:
    display_name: web-vm-1
    template: Linux Debian 7 64-bit
    service_offering: Tiny
    ssh_key: john@example.com
    tags:
      - { key: admin, value: john }
      {% if vm_group_tab is defined %}
      - { key: group, value: "{{ vm_group_tab }}" }
      {% endif %}
    user_data: |
      #cloud-config
      packages:
        - nginx

When vm_group_tab exists, the group tag gets added; when it doesn't, only the admin tag remains. It's easy to read and requires no extra tasks.

Method 2: Build Tags List with set_fact First #

If you prefer a cleaner task structure (or need to reuse the tags list elsewhere in your playbook), you can construct the dynamic tags list in a separate set_fact task:

- name: Build dynamic instance tags list
  set_fact:
    instance_tags:
      - { key: admin, value: john }
      {% if vm_group_tab is defined %}
      - { key: group, value: "{{ vm_group_tab }}" }
      {% endif %}

- name: create or update instance
  cs_instance:
    display_name: web-vm-1
    template: Linux Debian 7 64-bit
    service_offering: Tiny
    ssh_key: john@example.com
    tags: "{{ instance_tags }}"
    user_data: |
      #cloud-config
      packages:
        - nginx

This separates the tag logic from the instance creation task, making it easier to modify or reuse later.

Method 3: Filter with Ansible Native Filters #

For a more "Ansible-native" approach without Jinja2 conditionals, you can build a full list of potential tags and filter out the ones with undefined values:

- name: create or update instance
  cs_instance:
    display_name: web-vm-1
    template: Linux Debian 7 64-bit
    service_offering: Tiny
    ssh_key: john@example.com
    tags: "{{ [
        { key: 'admin', value: 'john' },
        { key: 'group', value: vm_group_tab | default(None) }
      ] | selectattr('value', 'defined') | list }}"
    user_data: |
      #cloud-config
      packages:
        - nginx

Here, we use default(None) to set the group tag's value to None if vm_group_tab is undefined, then selectattr('value', 'defined') filters out any tag where the value isn't present. This keeps the task self-contained without extra syntax blocks.

All three methods will give you exactly the behavior you want: the group tag is only included when vm_group_tab is defined.

内容的提问来源于stack exchange，提问作者nbari