feat: initial commit

tasks/ceph/enter-maint.yaml

@@ -0,0 +1,29 @@
+---
+
+# enable ceph maintenance mode
+- name: "enable ceph maintenance mode"
+  block:
+
+    # set ceph osd noout settings
+    - name: "set ceph noout"
+      ansible.builtin.shell:
+        cmd: "sudo ceph osd set noout"
+      changed_when: false
+      failed_when: "ceph_noout_result.rc != 0"
+      register: "ceph_noout_result"
+
+    # set ceph osd nobackfill settings
+    - name: "set ceph nobackfill"
+      ansible.builtin.shell:
+        cmd: "sudo ceph osd set nobackfill"
+      changed_when: false
+      failed_when: "ceph_nobackfill_result.rc != 0"
+      register: "ceph_nobackfill_result"
+
+    # set ceph osd norebalance settings
+    - name: "set ceph norebalance"
+      ansible.builtin.shell:
+        cmd: "sudo ceph osd set norebalance"
+      changed_when: false
+      failed_when: "ceph_norebalance_result.rc != 0"
+      register: "ceph_norebalance_result"

tasks/ceph/exit-maint.yaml

@@ -0,0 +1,43 @@
+---
+
+# exit ceph maintenance mode
+- name: "exit ceph maintenance mode"
+  block:
+
+    # unset ceph osd noout settings
+    - name: "unset ceph noout"
+      ansible.builtin.shell:
+        cmd: "sudo ceph osd unset noout"
+      changed_when: false
+      failed_when: "ceph_noout_result.rc != 0"
+      register: "ceph_noout_result"
+      tags: "cluster"
+    
+    # unset ceph osd nobackfill settings
+    - name: "unset ceph nobackfill"
+      ansible.builtin.shell:
+        cmd: "sudo ceph osd unset nobackfill"
+      changed_when: false
+      failed_when: "ceph_nobackfill_result.rc != 0"
+      register: "ceph_nobackfill_result"
+      tags: "cluster"
+    
+    # unset ceph osd norebalance settings
+    - name: "unset ceph norebalance"
+      ansible.builtin.shell:
+        cmd: "sudo ceph osd unset norebalance"
+      changed_when: false
+      failed_when: "ceph_norebalance_result.rc != 0"
+      register: "ceph_norebalance_result"
+      tags: "cluster"
+    
+    # wait for ceph to be healthy
+    - name: "wait for ceph to be healthy"
+      ansible.builtin.shell:
+        cmd: "sudo ceph -s"
+      changed_when: false
+      delay: 10
+      register: "ceph_status"
+      retries: 30
+      tags: "cluster"
+      until: "'HEALTH_OK' in ceph_status.stdout"

tasks/main.yaml

@@ -0,0 +1,25 @@
+---
+
+# update cluster
+- name: "update cluster"
+  tags: "update-cluster"
+  block:
+
+    # collect proxmox cluster nodes
+    - name: "collect proxmox cluster nodes"
+      set_fact:
+        leader_node: "{{ groups['proxmox'] | default([]) | first | default('') }}"
+
+    # enter ceph maintenance mode
+    - name: "set ceph enter maintenance mode"
+      ansible.builtin.include_tasks: "ceph/enter-maint.yaml"
+      when: "inventory_hostname == leader_node"
+
+    # update proxmox cluster
+    - name: "update proxmox cluster nodes"
+      ansible.builtin.include_tasks: "proxmox/update-node.yaml"
+
+    # exit ceph maintenance mode
+    - name: "exit ceph maintenance mode"
+      ansible.builtin.include_tasks: "ceph/exit-maint.yaml"
+      when: "inventory_hostname == leader_node"

tasks/proxmox/update-node.yaml

@@ -0,0 +1,64 @@
+---
+
+# enter maintenance mode
+- name: "enter maintenance mode"
+  ansible.builtin.shell:
+    cmd: "sudo /usr/sbin/ha-manager crm-command node-maintenance enable {{ inventory_hostname_short }}"
+  changed_when: false
+  failed_when: "maintenance_result.rc != 0"
+  register: "maintenance_result"
+
+# wait for host to be empty
+- name: "wait for host to be empty"
+  ansible.builtin.shell:
+    cmd: "set -o pipefail; \
+      sudo pvesh get /cluster/resources | \
+      grep {{ inventory_hostname_short }} | \
+      egrep -v \"node/pve|storage/pve|sdn/pve|template\""
+  changed_when: false
+  delay: 10
+  failed_when: "running_guests.rc >= 2"
+  register: "running_guests"
+  retries: 30
+  until: "running_guests.stdout_lines | length == 0"
+
+# install node updates
+- name: "install node updates"
+  ansible.builtin.apt:
+    dpkg_options: "force-confdef,force-confold"
+    force_apt_get: true
+    update_cache: true
+    upgrade: "dist"
+  changed_when: false
+
+# install firmware updates
+- name: "install firmware updates"
+  ansible.builtin.shell:
+    cmd: "sudo fwupdmgr upgrade --no-reboot-check --assume-yes"
+  changed_when: false
+  failed_when: false
+
+# initiate reboot
+- name: "reboot host"
+  ansible.builtin.reboot:
+    reboot_timeout: 300
+    pre_reboot_delay: 5
+    post_reboot_delay: 30
+    test_command: "uptime"
+  changed_when: false
+  when: "reboot_required"
+
+# exit maintenance mode
+- name: "exit maintenance mode"
+  ansible.builtin.shell:
+    cmd: "sudo /usr/sbin/ha-manager crm-command node-maintenance disable {{ inventory_hostname_short }}"
+  changed_when: false
+  delay: 10
+  failed_when: "maintenance_exit_result.rc != 0"
+  register: "maintenance_exit_result"
+  retries: 3
+
+# calm down period
+- name: "calming down"
+  ansible.builtin.wait_for:
+    timeout: 120