ansible_fencl/check_raid.yml

# check_raid.yml

- name: Check mdadm RAID status on VM (via SSH from proxmox_nextcloud)
  hosts: proxmox_nextcloud
  gather_facts: false
  become: true
  become_user: root
  become_method: sudo

  vars:
    # VM connection (provided by Semaphore env vars)
    vm_ip:   "{{ lookup('env', 'VM_IP') }}"
    vm_user: "{{ lookup('env', 'VM_USER') }}"
    vm_pass: "{{ lookup('env', 'VM_PASS') }}"
    use_sudo: false

    # Debug / behavior toggles (controlled via Semaphore variables)
    DEBUG: "{{ lookup('env', 'DEBUG') | default(0) | int }}"

    # Which md device to check (md0, md1, ... or /dev/md0)
    RAID_DEVICE: "{{ lookup('env', 'RAID_DEVICE') | default('md0', true) }}"

    # Fail the job if rebuild/resync is in progress (recommended)
    FAIL_ON_RESYNC: "{{ lookup('env', 'FAIL_ON_RESYNC') | default(1) | int }}"

    # Fail the job if a periodic "check" is running (usually set 0 to avoid noise)
    FAIL_ON_CHECK: "{{ lookup('env', 'FAIL_ON_CHECK') | default(0) | int }}"

  tasks:
    - name: Ensure sshpass is installed (for password-based SSH)  # English comments
      ansible.builtin.apt:
        name: sshpass
        state: present
        update_cache: yes
        cache_valid_time: 86400

    - name: Read /proc/mdstat from VM (via SSH)  # English comments
      ansible.builtin.command:
        argv:
          - timeout
          - 25s
          - sshpass
          - -e
          - ssh
          - -o
          - StrictHostKeyChecking=no
          - -o
          - UserKnownHostsFile=/dev/null
          - -o
          - ConnectTimeout=10
          - -o
          - ConnectionAttempts=1
          - -o
          - NumberOfPasswordPrompts=1
          - -o
          - PubkeyAuthentication=no
          - -o
          - GSSAPIAuthentication=no
          - -o
          - PasswordAuthentication=yes
          - -o
          - ServerAliveInterval=5
          - -o
          - ServerAliveCountMax=2
          - "{{ vm_user }}@{{ vm_ip }}"
          - "cat /proc/mdstat"
      environment:
        SSHPASS: "{{ vm_pass }}"
      register: mdstat_cmd
      changed_when: false
      failed_when: false
      no_log: "{{ DEBUG == 0 }}"

    - name: Fail if we cannot read /proc/mdstat (SSH/auth/network)  # English comments
      ansible.builtin.assert:
        that:
          - mdstat_cmd.rc == 0
        fail_msg: >-
          Cannot read /proc/mdstat from {{ vm_ip }} (rc={{ mdstat_cmd.rc }}).
          stderr={{ (mdstat_cmd.stderr | default('') | trim) }}
        success_msg: "Successfully read /proc/mdstat from {{ vm_ip }}."
      changed_when: false

    - name: Build base variables  # English comments
      ansible.builtin.set_fact:
        raid_md: "{{ RAID_DEVICE | regex_replace('^/dev/', '') }}"
        mdstat_text: "{{ mdstat_cmd.stdout | default('') }}"
      changed_when: false

    - name: Extract selected md block from mdstat  # English comments
      ansible.builtin.set_fact:
        raid_block: >-
          {{
            (
              mdstat_text
              | regex_findall(
                  '(?ms)^' ~ raid_md ~ '\\s*:.*?(?=^md\\d+\\s*:|^unused devices:|\\Z)'
                )
              | first
              | default('', true)
            )
          }}
      changed_when: false

    - name: Parse RAID status from mdstat (safe parsing)  # English comments
      ansible.builtin.set_fact:
        raid_present: "{{ (raid_block | length) > 0 }}"
        raid_status: >-
          {{
            (
              raid_block
              | regex_findall('\\[[0-9]+/[0-9]+\\]\\s*\\[([U_]+)\\]')
              | first
              | default('', true)
            )
          }}
        raid_is_degraded: "{{ (raid_status | length > 0) and ('_' in raid_status) }}"
        raid_is_rebuilding: "{{ raid_block is search('(?i)\\b(resync|recovery|reshape|repair)\\b') }}"
        raid_is_checking: "{{ raid_block is search('(?i)\\bcheck\\b') }}"
        raid_action_line: >-
          {{
            (
              raid_block
              | regex_findall('(?im)^(\\s*\\[[^\\]]+\\].*\\b(?:resync|recovery|reshape|repair|check)\\b.*)$')
              | first
              | default('', true)
            )
          }}
        raid_progress: >-
          {{
            (
              raid_block
              | regex_findall('(?i)\\b(?:resync|recovery|reshape|repair|check)\\b\\s*=\\s*([0-9.]+)%')
              | first
              | default('', true)
            )
          }}
      changed_when: false

    - name: Debug | Show mdstat and parsed values  # English comments
      ansible.builtin.debug:
        msg: |
          --- /proc/mdstat ---
          {{ mdstat_text }}
          --- md block ({{ raid_md }}) ---
          {{ raid_block }}
          --- Parsed ---
          raid_present={{ raid_present }}
          raid_status={{ raid_status }}
          raid_is_degraded={{ raid_is_degraded }}
          raid_is_rebuilding={{ raid_is_rebuilding }}
          raid_is_checking={{ raid_is_checking }}
          raid_progress={{ raid_progress }}
          raid_action_line={{ raid_action_line }}
      when: DEBUG == 1

    - name: Fail if RAID device is not present  # English comments
      ansible.builtin.assert:
        that:
          - raid_present
        fail_msg: "RAID {{ raid_md }} not found in /proc/mdstat on VM {{ vm_ip }}."
        success_msg: "RAID {{ raid_md }} found in /proc/mdstat."
      changed_when: false

    - name: Fail if RAID is degraded (missing member)  # English comments
      ansible.builtin.assert:
        that:
          - not raid_is_degraded
        fail_msg: >-
          RAID {{ raid_md }} is DEGRADED: status={{ raid_status }}.
          mdstat excerpt={{ (raid_block | trim) }}
        success_msg: "RAID {{ raid_md }} is OK: status={{ raid_status }}."
      changed_when: false

    - name: Fail if RAID rebuild/resync is in progress (optional)  # English comments
      ansible.builtin.assert:
        that:
          - not raid_is_rebuilding
        fail_msg: >-
          RAID {{ raid_md }} is rebuilding/resyncing.
          {{ 'progress=' ~ raid_progress ~ '%; ' if (raid_progress | length) > 0 else '' }}
          line={{ (raid_action_line | default('n/a', true)) }}
        success_msg: "RAID {{ raid_md }} is not rebuilding/resyncing."
      when: FAIL_ON_RESYNC | int == 1
      changed_when: false

    - name: Fail if RAID check is in progress (optional)  # English comments
      ansible.builtin.assert:
        that:
          - not raid_is_checking
        fail_msg: >-
          RAID {{ raid_md }} is running a periodic check.
          {{ 'progress=' ~ raid_progress ~ '%; ' if (raid_progress | length) > 0 else '' }}
          line={{ (raid_action_line | default('n/a', true)) }}
        success_msg: "RAID {{ raid_md }} is not running a periodic check."
      when: FAIL_ON_CHECK | int == 1
      changed_when: false