---

- name: Managed Lab Maintenance
  hosts: all_nodes
  become: true
  serial: 2  # Update 2 nodes at a time to reduce network/storage strain

  tasks:
    - name: 1. Update all packages via Nala
      ansible.builtin.apt:
        update_cache: yes
        upgrade: dist
        autoremove: yes
      register: apt_res

    - name: 2. Check if a reboot is required
      ansible.builtin.stat:
        path: /var/run/reboot-required
      register: reboot_required_file

    - name: 3. Conditional Reboot (Skip Physical Hosts for now)
      when: 
        - reboot_required_file.stat.exists
        - inventory_hostname not in groups['physical_iron']  # Handle physical hosts in a separate play
      ansible.builtin.reboot:
        msg: "Rebooting {{ inventory_hostname }} after updates"
        reboot_timeout: 600

- name: Final Tier - Proxmox Host Maintenance
  hosts: physical_iron
  become: true
  serial: 1  # Reboot physical hosts sequentially to avoid complete cluster downtime
  
  tasks:
    - name: Check if Host needs reboot
      ansible.builtin.stat:
        path: /var/run/reboot-required
      register: host_reboot

    - name: Proxmox Host Reboot Block
      when: host_reboot.stat.exists
      block:
        - name: Graceful shutdown of all VMs/LXCs
          ansible.builtin.shell: pvesh create /nodes/localhost/stopall
          failed_when: false

        - name: Reboot the Host
          ansible.builtin.reboot:
            msg: "Host maintenance reboot"
            reboot_timeout: 900