testing new proxmox logic

roles/proxmox_ceph/defaults/main.yml

@@ -0,0 +1,18 @@
+---
+# =============================================================================
+# proxmox_ceph — defaults
+# =============================================================================
+
+# Action: set_noout | clear_noout | status | check_health
+ceph_action: status
+
+# Health check settings
+ceph_health_timeout: 300        # seconds to wait for HEALTH_OK
+ceph_health_retries: 30
+ceph_health_delay: 10
+
+# Abort upgrade if CEPH is in error state
+ceph_abort_on_error: true
+
+# Warn but continue if CEPH has warnings
+ceph_warn_on_warning: true

roles/proxmox_ceph/meta/main.yml

@@ -0,0 +1,11 @@
+---
+galaxy_info:
+  role_name: proxmox_ceph
+  author: ansible-msp
+  description: "MSP Proxmox automation — proxmox_ceph"
+  min_ansible_version: "2.15"
+  platforms:
+    - name: Debian
+      versions:
+        - bookworm
+dependencies: []

roles/proxmox_ceph/tasks/main.yml

@@ -0,0 +1,140 @@
+---
+# =============================================================================
+# proxmox_ceph — tasks
+# Manages CEPH noout flag and health checks during maintenance.
+# Skips gracefully if CEPH is not configured on this cluster.
+# =============================================================================
+
+# ── Detect CEPH ───────────────────────────────────────────────────────────────
+- name: "CEPH | Detect if CEPH is configured"
+  ansible.builtin.command: ceph status
+  register: ceph_detect
+  changed_when: false
+  failed_when: false
+  run_once: true
+
+- name: "CEPH | Set CEPH enabled fact"
+  ansible.builtin.set_fact:
+    ceph_is_enabled: "{{ ceph_detect.rc == 0 }}"
+  run_once: true
+
+- name: "CEPH | Skip — CEPH not configured on this cluster"
+  ansible.builtin.debug:
+    msg: "CEPH is not configured on this cluster — skipping all CEPH tasks."
+  when: not ceph_is_enabled
+  run_once: true
+
+# ── CEPH status ───────────────────────────────────────────────────────────────
+- name: "CEPH | Get cluster status"
+  ansible.builtin.command: ceph status --format json
+  register: ceph_status_raw
+  changed_when: false
+  run_once: true
+  when: ceph_is_enabled
+
+- name: "CEPH | Parse status"
+  ansible.builtin.set_fact:
+    ceph_status: "{{ ceph_status_raw.stdout | from_json }}"
+  run_once: true
+  when: ceph_is_enabled
+
+- name: "CEPH | Log health"
+  ansible.builtin.debug:
+    msg: "CEPH health: {{ ceph_status.health.status }}"
+  run_once: true
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'status'
+
+# ── Health check ──────────────────────────────────────────────────────────────
+- name: "CEPH | Check health | Abort if HEALTH_ERR"
+  ansible.builtin.fail:
+    msg: >-
+      CEPH is in HEALTH_ERR state — aborting to prevent data loss.
+      Run 'ceph status' to investigate. Set ceph_abort_on_error=false to override.
+  when:
+    - ceph_is_enabled
+    - ceph_abort_on_error
+    - ceph_status.health.status == 'HEALTH_ERR'
+  run_once: true
+
+- name: "CEPH | Check health | Warn on HEALTH_WARN"
+  ansible.builtin.debug:
+    msg: >-
+      WARNING: CEPH is in HEALTH_WARN state.
+      Proceeding — set ceph_warn_on_warning=false to suppress this message.
+      Checks: {{ ceph_status.health.checks | default({}) | dict2items | map(attribute='key') | list }}
+  when:
+    - ceph_is_enabled
+    - ceph_warn_on_warning
+    - ceph_status.health.status == 'HEALTH_WARN'
+  run_once: true
+
+# ── Set noout ─────────────────────────────────────────────────────────────────
+- name: "CEPH | Set noout flag"
+  ansible.builtin.command: ceph osd set noout
+  changed_when: true
+  run_once: true
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'set_noout'
+
+- name: "CEPH | Confirm noout set"
+  ansible.builtin.debug:
+    msg: "✓ CEPH noout flag SET — OSDs will not be marked out during maintenance."
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'set_noout'
+  run_once: true
+
+# ── Clear noout ───────────────────────────────────────────────────────────────
+- name: "CEPH | Clear noout flag"
+  ansible.builtin.command: ceph osd unset noout
+  changed_when: true
+  run_once: true
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'clear_noout'
+
+- name: "CEPH | Wait for HEALTH_OK after clearing noout"
+  ansible.builtin.command: ceph status --format json
+  register: ceph_recovery_check
+  changed_when: false
+  until: "(ceph_recovery_check.stdout | from_json).health.status in ['HEALTH_OK', 'HEALTH_WARN']"
+  retries: "{{ ceph_health_retries }}"
+  delay: "{{ ceph_health_delay }}"
+  run_once: true
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'clear_noout'
+
+- name: "CEPH | Log recovery status"
+  ansible.builtin.debug:
+    msg: >-
+      ✓ CEPH noout CLEARED —
+      health: {{ (ceph_recovery_check.stdout | from_json).health.status }}
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'clear_noout'
+  run_once: true
+
+# ── check_health action ───────────────────────────────────────────────────────
+- name: "CEPH | Wait for healthy state"
+  ansible.builtin.command: ceph status --format json
+  register: ceph_health_wait
+  changed_when: false
+  until: "(ceph_health_wait.stdout | from_json).health.status in ['HEALTH_OK', 'HEALTH_WARN']"
+  retries: "{{ ceph_health_retries }}"
+  delay: "{{ ceph_health_delay }}"
+  run_once: true
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'check_health'
+
+- name: "CEPH | Health check result"
+  ansible.builtin.debug:
+    msg: "CEPH health: {{ (ceph_health_wait.stdout | from_json).health.status }}"
+  when:
+    - ceph_is_enabled
+    - ceph_action == 'check_health'
+  run_once: true