Ansible运行时环境变量及运行环境相关技术咨询

Hey there! Let's tackle your two Ansible questions with practical, actionable advice:

Ansible interacts with environment variables at both the control node (your local machine) and managed nodes (target servers) — here's the breakdown:

• Control node environment variables: Ansible inherits your shell's environment by default. Key ones to know include:
  • ANSIBLE_CONFIG: Points to a custom Ansible config file (instead of the default /etc/ansible/ansible.cfg).
  • ANSIBLE_ROLES_PATH: Adds custom directories to Ansible's role search path.
  • ANSIBLE_PYTHON_INTERPRETER: Sets the default Python interpreter Ansible uses on the control node.
  • PATH: Critical for finding system tools (like ssh, mysql) that Ansible relies on.
• Managed node environment variables: By default, Ansible runs tasks in a non-login shell, so it won't load profile files like ~/.bash_profile or /etc/profile. If your tasks need specific env vars (e.g., a custom PATH for tools), you have two options:
  1. Use the environment keyword in your playbook to define variables per task or play.
  2. Explicitly source profile files in your command (e.g., source ~/.bashrc && my_command).
• Virtualenv integration: When you activate a virtualenv on the control node, Ansible will inherit that env's variables (like PYTHONPATH and the active Python interpreter) — this is why many users use virtualenvs to isolate Ansible dependencies.

It sounds like you're hitting common environment mismatch and dependency gaps — let's walk through fixes tailored to your scenario:

• Align Python interpreters across nodes: A frequent culprit is mismatched Python versions between your control node (virtualenv) and managed nodes. For example, if your virtualenv uses Python 3.9 but managed nodes default to Python 2.7, modules like mysql will fail. Fix this by explicitly setting the interpreter in your playbook:

  - hosts: your_targets
    vars:
      # Use the path to Python 3 on your managed nodes (or their virtualenv path)
      ansible_python_interpreter: /usr/bin/python3
    tasks:
      # Your module tasks here
  

• Ensure full dependency installation: Modules like mysql_db or mysql_user require both Python packages and system-level libraries:
  • On the control node (virtualenv): Run pip install pymysql mysqlclient to install the required Python drivers.
  • On managed nodes: Install system libraries (e.g., libmysqlclient-dev on Ubuntu, mysql-devel on RHEL/CentOS) and the same Python drivers if the module runs locally on the managed node.
• Eliminate Ansible version conflicts: If you have multiple Ansible installations (e.g., system-wide and virtualenv), you might be accidentally running the wrong one. Verify with ansible --version — check that the Python interpreter path matches your virtualenv. If not, use the full path to your virtualenv's Ansible binary (e.g., ~/my_venv/bin/ansible-playbook) instead of just ansible-playbook.
• Fix "file not found" errors on managed nodes: This almost always relates to a restricted PATH in the non-login shell. Extend the path for your task using the environment keyword:

  - name: Run a tool installed in /usr/local/bin
    command: my_custom_tool
    environment:
      PATH: "/usr/local/bin:{{ ansible_env.PATH }}"
  

• Adapt official examples to your environment: Official docs sometimes assume default setups (e.g., Python 2 or legacy dependencies). For MySQL modules, make sure you're using PyMySQL (Python 3 compatible) and explicitly define connection parameters (like login_host, login_user) instead of relying on defaults that might not match your setup.

内容的提问来源于stack exchange，提问作者New2Python