声明式管理 Proxmox API 对象与 Samba DNS 上游

infrastructure/proxmox/ansible/group_vars/pve.yml

@@ -1,4 +1,13 @@
 ---
+# Shared authentication for typed community.proxmox API modules. Secrets are
+# runtime inputs; never put them in inventory or vars files.
+pve_api_host: "{{ lookup('env', 'PROXMOX_HOST') }}"
+pve_api_user: "{{ lookup('env', 'PROXMOX_USER') }}"
+pve_api_token_id: "{{ lookup('env', 'PROXMOX_TOKEN_ID') }}"
+pve_api_token_secret: "{{ lookup('env', 'PROXMOX_TOKEN_SECRET') }}"
+pve_api_validate_certs: true
+pve_api_python_interpreter: "{{ lookup('env', 'PVE_API_PYTHON') | default(ansible_playbook_python, true) }}"
+
 ansible_user: root
 ansible_python_interpreter: /usr/bin/python3
 

infrastructure/proxmox/ansible/inventory/hosts.yml

@@ -19,20 +19,13 @@ all:
 
         pve2:
           ansible_host: 192.168.10.7
-          # ThinkCentre, Ryzen 2400GE (Raven Ridge), 8G. THIS is the node that
+          # ThinkCentre, Ryzen 2400GE (Raven Ridge), 8G.
-          # randomly froze. Kernel workaround below; also set BIOS
-          # "Power Supply Idle Control" -> "Typical Current Idle" (the more
-          # reliable half of the fix, and it can only be done physically).
-          pve_kernel_cmdline_extra: "idle=nomwait processor.max_cstate=1"
           pve_install_disk: nvme0n1
           pve_bulk_disk: sda
 
         pve3:
           ansible_host: 192.168.10.9
-          # Same Raven Ridge silicon as pve2 but has never frozen. Deliberately
+          # ThinkCentre, Ryzen 2400GE (Raven Ridge), 8G.
-          # left WITHOUT the C-state workaround: processor.max_cstate=1 blocks
-          # deeper idle states and costs power, so it is not applied pre-emptively.
-          # If pve3 ever freezes, set pve_kernel_cmdline_extra here too.
           pve_install_disk: nvme0n1
           pve_bulk_disk: sda
 

infrastructure/proxmox/ansible/requirements.txt

@@ -0,0 +1,2 @@
+proxmoxer==2.3.0
+requests==2.34.2

infrastructure/proxmox/ansible/requirements.yml

@@ -13,3 +13,5 @@ collections:
   - name: ansible.posix         # mount, authorized_key
   - name: community.general     # proxmox_kvm (guest lifecycle). NOTE: deb822_repository
                                 # is ansible.builtin, NOT community.general — see CLAUDE.md
+  - name: community.proxmox
+    version: 2.0.0              # typed PVE API modules for HA, SDN and guest lifecycle

infrastructure/proxmox/ansible/roles/pve_ha/defaults/main.yml

@@ -14,12 +14,26 @@ pve_ha_resources:
     max_restart: 3
     max_relocate: 2
 
+# pve3 is the normal home for the critical router. Non-strict preserves pve1
+# and pve2 as failover targets when pve3 is unavailable.
+pve_ha_rules:
+  - name: prefer-vyos-pve3
+    type: node-affinity
+    resources: ["vm:100"]
+    nodes: ["pve3:100", "pve1:10", "pve2:10"]
+    strict: false
+    comment: "Prefer critical VyOS router on pve3, away from CI nodes"
+
+pve_ha_portable_vms:
+  - vmid: 100
+    node: pve3
+    cpu: x86-64-v2-AES
+
 # --- Watchdog ---------------------------------------------------------------
 # ⚠ WHY THIS MATTERS: adding ANY HA resource arms fencing cluster-wide. PVE's
 # default is `softdog`, a SOFTWARE watchdog — a kernel timer, which CANNOT fire if
-# the kernel itself is frozen. That is precisely the failure this cluster has
+# the kernel itself is frozen. A hardware watchdog is independent silicon and
-# actually seen (pve2's Raven Ridge idle freeze). A hardware watchdog is
+# fires regardless.
-# independent silicon and fires regardless.
 #
 # Verified available 2026-07-26:
 #   pve1 (Intel i3-6100U) -> iTCO_wdt      (timeout 30s)

infrastructure/proxmox/ansible/roles/pve_ha/handlers/main.yml

@@ -2,7 +2,7 @@
 # Deliberately does NOT reboot. Swapping the watchdog under an armed cluster is a
 # manual, ordered operation (see ../../../README-ha.md); a surprise rolling reboot
 # of all three nodes is exactly what you do not want an idempotent play to do.
-- name: reboot required for watchdog
+- name: Reboot required for watchdog
   ansible.builtin.debug:
     msg: >-
       Watchdog config changed on {{ inventory_hostname }}. It takes effect on the

infrastructure/proxmox/ansible/roles/pve_ha/tasks/main.yml

@@ -1,10 +1,19 @@
 ---
-# Proxmox HA. Read roles/pve_ha/defaults/main.yml first — it explains why the
+- name: Validate Proxmox API inputs
-# resource list is short and why the watchdog choice is not cosmetic.
+  ansible.builtin.assert:
+    that:
+      - pve_api_host | length > 0
+      - pve_api_user | length > 0
+      - pve_api_token_id | length > 0
+      - pve_api_token_secret | length > 0
+    fail_msg: >-
+      Export PROXMOX_HOST, PROXMOX_USER, PROXMOX_TOKEN_ID and
+      PROXMOX_TOKEN_SECRET before running ha.yml.
+  run_once: true
+  delegate_to: localhost
+  no_log: true
+  tags: [ha]
 
-# --- Watchdog ---------------------------------------------------------------
-# Done BEFORE registering resources: adding a resource arms fencing, and arming
-# fencing on a software watchdog is the weakest configuration.
 - name: Select the hardware watchdog module
   ansible.builtin.lineinfile:
     path: /etc/default/pve-ha-manager
@@ -12,13 +21,9 @@
     line: "WATCHDOG_MODULE={{ pve_ha_watchdog_module }}"
     create: false
   when: pve_ha_watchdog_module | length > 0
-  notify: reboot required for watchdog
+  notify: Reboot required for watchdog
   tags: [ha, watchdog]
 
-# watchdog-mux opens /dev/watchdog, which belongs to whichever watchdog registered
-# FIRST. PVE loads softdog at boot, so it wins and the hardware module ends up as
-# an unused watchdog1. Blacklisting softdog is what actually makes the hardware
-# one take effect — setting WATCHDOG_MODULE alone does nothing.
 - name: Blacklist softdog so the hardware watchdog claims /dev/watchdog
   ansible.builtin.copy:
     dest: /etc/modprobe.d/pve-ha-watchdog.conf
@@ -29,48 +34,147 @@
       # {{ pve_ha_watchdog_module }} register as watchdog0 and own /dev/watchdog.
       blacklist softdog
   when: pve_ha_watchdog_module | length > 0
-  notify: reboot required for watchdog
+  notify: Reboot required for watchdog
   tags: [ha, watchdog]
 
-# ⚠ NOT DONE LIVE ON PURPOSE. Swapping the watchdog on a running node means
+- name: Read current HA VM configurations
-# stopping watchdog-mux and unloading softdog while fencing is armed — get the
+  ansible.builtin.uri:
-# order wrong and the node self-fences (reboots). The safe live procedure is
+    url: >-
-# documented in ../../README-ha.md; otherwise it simply takes effect on the next
+      https://{{ pve_api_host }}:8006/api2/json/nodes/{{ item.node }}/qemu/{{ item.vmid }}/config
-# reboot, which is why the handler only WARNS rather than rebooting anything.
+    headers:
-
+      Authorization: >-
-# --- Resources ---------------------------------------------------------------
+        PVEAPIToken={{ pve_api_user }}!{{ pve_api_token_id }}={{ pve_api_token_secret }}
-- name: Read current HA resources
+    ca_path: /etc/ssl/certs/ca-certificates.crt
-  ansible.builtin.command: ha-manager status
+    return_content: true
-  register: pve_ha_status
+  loop: "{{ pve_ha_portable_vms }}"
-  changed_when: false
+  loop_control:
-  failed_when: false
+    label: "VM {{ item.vmid }}"
+  register: pve_ha_vm_config
   run_once: true
+  delegate_to: localhost
+  no_log: true
   tags: [ha]
 
-- name: Register HA resources
+- name: Reconcile portable CPU models for HA VMs
-  ansible.builtin.command: >-
+  community.proxmox.proxmox_kvm:
-    ha-manager add {{ item.sid }}
+    api_host: "{{ pve_api_host }}"
-    --state {{ item.state }}
+    api_user: "{{ pve_api_user }}"
-    --max_restart {{ item.max_restart }}
+    api_token_id: "{{ pve_api_token_id }}"
-    --max_relocate {{ item.max_relocate }}
+    api_token_secret: "{{ pve_api_token_secret }}"
+    validate_certs: "{{ pve_api_validate_certs }}"
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+    node: "{{ item.node }}"
+    vmid: "{{ item.vmid }}"
+    cpu: "{{ item.cpu }}"
+    update: true
+    state: present
+  loop: "{{ pve_ha_portable_vms }}"
+  loop_control:
+    label: "VM {{ item.vmid }} -> {{ item.cpu }}"
+    index_var: vm_index
+  when: pve_ha_vm_config.results[vm_index].json.data.cpu != item.cpu
+  run_once: true
+  delegate_to: localhost
+  no_log: true
+  vars:
+    ansible_python_interpreter: "{{ pve_api_python_interpreter }}"
+  tags: [ha]
+
+- name: Read current HA resources
+  ansible.builtin.uri:
+    url: "https://{{ pve_api_host }}:8006/api2/json/cluster/ha/resources"
+    headers:
+      Authorization: >-
+        PVEAPIToken={{ pve_api_user }}!{{ pve_api_token_id }}={{ pve_api_token_secret }}
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+    return_content: true
+  register: pve_ha_resource_state
+  run_once: true
+  delegate_to: localhost
+  no_log: true
+  tags: [ha]
+
+- name: Reconcile HA resources through the PVE API
+  community.proxmox.proxmox_cluster_ha_resources:
+    api_host: "{{ pve_api_host }}"
+    api_user: "{{ pve_api_user }}"
+    api_token_id: "{{ pve_api_token_id }}"
+    api_token_secret: "{{ pve_api_token_secret }}"
+    validate_certs: "{{ pve_api_validate_certs }}"
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+    name: "{{ item.sid }}"
+    state: present
+    hastate: "{{ item.state }}"
+    max_restart: "{{ item.max_restart }}"
+    max_relocate: "{{ item.max_relocate }}"
   loop: "{{ pve_ha_resources }}"
   loop_control:
     label: "{{ item.sid }}"
-  # `ha-manager add` errors if the resource already exists, so gate on the status
+  when: >-
-  # output. This is what keeps a second run at changed=0.
+    _current_resource | length == 0 or
-  when: "'service ' ~ item.sid not in (pve_ha_status.stdout | default(''))"
+    _current_resource.state != item.state or
+    _current_resource.max_restart | int != item.max_restart | int or
+    _current_resource.max_relocate | int != item.max_relocate | int
   run_once: true
+  delegate_to: localhost
+  no_log: true
+  vars:
+    ansible_python_interpreter: "{{ pve_api_python_interpreter }}"
+    _current_resource: >-
+      {{ pve_ha_resource_state.json.data |
+         selectattr('sid', 'equalto', item.sid) | first | default({}) }}
   tags: [ha]
 
-- name: Verify HA is armed and the resource is known
+- name: Reconcile HA placement rules through the PVE API
-  ansible.builtin.command: ha-manager status
+  community.proxmox.proxmox_cluster_ha_rules:
-  register: pve_ha_verify
+    api_host: "{{ pve_api_host }}"
-  changed_when: false
+    api_user: "{{ pve_api_user }}"
+    api_token_id: "{{ pve_api_token_id }}"
+    api_token_secret: "{{ pve_api_token_secret }}"
+    validate_certs: "{{ pve_api_validate_certs }}"
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+    name: "{{ item.name }}"
+    state: present
+    type: "{{ item.type }}"
+    resources: "{{ item.resources }}"
+    nodes: "{{ item.nodes }}"
+    strict: "{{ item.strict }}"
+    comment: "{{ item.comment }}"
+    disable: false
+  loop: "{{ pve_ha_rules }}"
+  loop_control:
+    label: "{{ item.name }}"
   run_once: true
+  delegate_to: localhost
+  no_log: true
+  vars:
+    ansible_python_interpreter: "{{ pve_api_python_interpreter }}"
   tags: [ha]
 
-- name: Show HA state
+- name: Read HA placement rules through the PVE API
-  ansible.builtin.debug:
+  community.proxmox.proxmox_cluster_ha_rules_info:
-    msg: "{{ pve_ha_verify.stdout_lines }}"
+    api_host: "{{ pve_api_host }}"
+    api_user: "{{ pve_api_user }}"
+    api_token_id: "{{ pve_api_token_id }}"
+    api_token_secret: "{{ pve_api_token_secret }}"
+    validate_certs: "{{ pve_api_validate_certs }}"
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+  register: pve_ha_rule_state
   run_once: true
+  delegate_to: localhost
+  no_log: true
+  vars:
+    ansible_python_interpreter: "{{ pve_api_python_interpreter }}"
+  tags: [ha]
+
+- name: Assert declared placement rules exist
+  ansible.builtin.assert:
+    that:
+      - >-
+        (pve_ha_rules | map(attribute='name') | list |
+         difference(pve_ha_rule_state.rules | default([]) |
+                    map(attribute='rule') | list) | length) == 0
+    fail_msg: A declared HA placement rule is absent after reconciliation
+  run_once: true
+  delegate_to: localhost
   tags: [ha]

infrastructure/proxmox/ansible/roles/pve_kernel_params/tasks/main.yml

@@ -1,7 +1,6 @@
 ---
 # Per-host kernel cmdline additions, driven by `pve_kernel_cmdline_extra` in the
-# inventory. Currently only pve2 sets it (Raven Ridge random-freeze workaround);
+# inventory. No node currently sets it; keep the role for future kernel parameters.
-# pve3 is identical silicon but is deliberately left alone -- see inventory.
 #
 # This is a no-op on hosts that do not define the variable.
 

infrastructure/proxmox/ansible/roles/pve_sdn/tasks/main.yml

@@ -1,60 +1,81 @@
 ---
-# SDN config is cluster-wide (/etc/pve/sdn/), so everything runs once.
+- name: Validate Proxmox API inputs
-
+  ansible.builtin.assert:
-- name: Read existing zones
+    that:
-  ansible.builtin.command:
+      - pve_api_host | length > 0
-    cmd: "pvesh get /cluster/sdn/zones --output-format json"
+      - pve_api_user | length > 0
-  register: _zones
+      - pve_api_token_id | length > 0
-  changed_when: false
+      - pve_api_token_secret | length > 0
-  check_mode: false
+    fail_msg: Missing Proxmox API environment variables
   run_once: true
+  delegate_to: localhost
+  no_log: true
 
-- name: Create the VLAN zone
+- name: Read current PVE SDN zones
-  ansible.builtin.command:
+  ansible.builtin.uri:
-    cmd: >-
+    url: "https://{{ pve_api_host }}:8006/api2/json/cluster/sdn/zones"
-      pvesh create /cluster/sdn/zones --zone {{ pve_sdn_zone }} --type vlan
+    headers:
-      --bridge {{ pve_sdn_bridge }} --mtu {{ pve_sdn_mtu }}
+      Authorization: >-
-      --nodes {{ groups['pve'] | join(',') }}
+        PVEAPIToken={{ pve_api_user }}!{{ pve_api_token_id }}={{ pve_api_token_secret }}
-  when: pve_sdn_zone not in (_zones.stdout | from_json | map(attribute='zone') | list)
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+    return_content: true
+  register: pve_sdn_zone_state
   run_once: true
+  delegate_to: localhost
+  no_log: true
 
-- name: Read existing vnets
+- name: Reconcile the PVE VLAN zone
-  ansible.builtin.command:
+  community.proxmox.proxmox_zone:
-    cmd: "pvesh get /cluster/sdn/vnets --output-format json"
+    api_host: "{{ pve_api_host }}"
-  register: _vnets
+    api_user: "{{ pve_api_user }}"
-  changed_when: false
+    api_token_id: "{{ pve_api_token_id }}"
-  check_mode: false
+    api_token_secret: "{{ pve_api_token_secret }}"
+    validate_certs: "{{ pve_api_validate_certs }}"
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+    zone: "{{ pve_sdn_zone }}"
+    type: vlan
+    bridge: "{{ pve_sdn_bridge }}"
+    mtu: "{{ pve_sdn_mtu }}"
+    nodes: "{{ groups['pve'] | join(',') }}"
+    update: true
+    state: present
+  when: >-
+    _current_zone | length == 0 or
+    _current_zone.type != 'vlan' or
+    _current_zone.bridge != pve_sdn_bridge or
+    _current_zone.mtu | int != pve_sdn_mtu | int or
+    _current_zone.nodes != (groups['pve'] | join(','))
   run_once: true
+  delegate_to: localhost
+  no_log: true
+  vars:
+    ansible_python_interpreter: "{{ pve_api_python_interpreter }}"
+    _current_zone: >-
+      {{ pve_sdn_zone_state.json.data |
+         selectattr('zone', 'equalto', pve_sdn_zone) | first | default({}) }}
 
-- name: Create the VNets
+# The module owns the SDN global lock and applies pending configuration before
-  ansible.builtin.command:
+# releasing it; a separate unconditional `pvesh set /cluster/sdn` is harmful
-    cmd: >-
+# because it reports changed on every run.
-      pvesh create /cluster/sdn/vnets --vnet {{ item.name }}
+- name: Reconcile PVE VNets
-      --zone {{ pve_sdn_zone }} --tag {{ item.tag }}
+  community.proxmox.proxmox_vnet:
-      --alias '{{ item.alias }}'
+    api_host: "{{ pve_api_host }}"
+    api_user: "{{ pve_api_user }}"
+    api_token_id: "{{ pve_api_token_id }}"
+    api_token_secret: "{{ pve_api_token_secret }}"
+    validate_certs: "{{ pve_api_validate_certs }}"
+    ca_path: /etc/ssl/certs/ca-certificates.crt
+    vnet: "{{ item.name }}"
+    zone: "{{ pve_sdn_zone }}"
+    tag: "{{ item.tag }}"
+    alias: "{{ item.alias }}"
+    update: true
+    state: present
   loop: "{{ pve_sdn_vnets }}"
   loop_control:
-    label: "{{ item.name }} (vlan {{ item.tag }})"
+    label: "{{ item.name }} (VLAN {{ item.tag }})"
-  when: item.name not in (_vnets.stdout | from_json | map(attribute='vnet') | list)
-  run_once: true
-
-- name: Apply the SDN configuration
-  # SDN changes stay PENDING until applied; without this the VNet bridges are
-  # never actually created on the nodes.
-  ansible.builtin.command:
-    cmd: "pvesh set /cluster/sdn"
-  register: _apply
-  changed_when: true
-  run_once: true
-
-- name: Report
-  ansible.builtin.shell:
-    cmd: "pvesh get /cluster/sdn/vnets --output-format json | python3 -c \"import json,sys;[print('   ',v['vnet'],'vlan',v.get('tag'),'zone',v.get('zone')) for v in json.load(sys.stdin)]\""
-  register: _rep
-  changed_when: false
-  run_once: true
-
-- name: Show it
-  ansible.builtin.debug:
-    msg: "{{ _rep.stdout_lines }}"
   run_once: true
+  delegate_to: localhost
+  no_log: true
+  vars:
+    ansible_python_interpreter: "{{ pve_api_python_interpreter }}"

infrastructure/proxmox/ansible/roles/pve_vm/defaults/main.yml

@@ -28,6 +28,9 @@ pve_vms:
     description: "VyOS router: OSPF peer to the NEC IX, gateway for the SDN VNets."
     cores: 2
     memory: 2048
+    # Portable across Intel pve1 and AMD pve2/pve3. `host` transferred memory
+    # but failed target-side resume during the 2026-09-17 migration.
+    cpu: x86-64-v2-AES
     disk: "pve-rg:8"
     # Headless by design: UEFI + serial console, NO emulated VGA. VyOS's own
     # kernel cmdline carries console=ttyS0, and the installer was answered with
@@ -56,6 +59,7 @@ pve_vms:
     # are AMD.
     cores: 4
     memory: 4096
+    cpu: host
     disk: "pve-rg:40"
     bios: ovmf
     machine: q35

infrastructure/proxmox/ansible/roles/pve_vm/tasks/main.yml

@@ -40,7 +40,7 @@
       --description '{{ item.item.description }}'
       --machine {{ item.item.machine | default('q35') }}
       --bios {{ item.item.bios | default('ovmf') }}
-      --memory {{ item.item.memory }} --cores {{ item.item.cores }} --cpu host
+      --memory {{ item.item.memory }} --cores {{ item.item.cores }} --cpu {{ item.item.cpu | default('x86-64-v2-AES') }}
       --scsihw virtio-scsi-single
       --scsi0 {{ item.item.disk }},discard=on,ssd=1
       --efidisk0 {{ item.item.disk.split(':')[0] }}:1,efitype=4m,pre-enrolled-keys=0

infrastructure/proxmox/ansible/roles/vyos_router/defaults/main.yml

@@ -71,3 +71,9 @@ vyos_ospf_router_id: "192.168.10.2"
 vyos_ospf_area: "0"
 vyos_ospf_networks:
   - "192.168.10.0/24"
+
+# Stable K3s datastore endpoint. Only the declared writable primary is present
+# in the backend; PostgreSQL health checks must never be treated as promotion.
+vyos_postgresql_listen_address: "10.60.0.1"
+vyos_postgresql_port: 5432
+vyos_postgresql_primary_address: "10.60.0.11"

infrastructure/proxmox/ansible/roles/vyos_router/templates/vyos.conf.j2

@@ -29,6 +29,17 @@ set protocols ospf area {{ vyos_ospf_area }} network {{ i.network }}
 set protocols ospf interface {{ i.iface }} passive
 {% endfor %}
 
+{# --- K3s PostgreSQL datastore: fixed endpoint, explicitly selected primary --- #}
+set load-balancing haproxy service sandbox-postgresql listen-address {{ vyos_postgresql_listen_address }}
+set load-balancing haproxy service sandbox-postgresql port {{ vyos_postgresql_port }}
+set load-balancing haproxy service sandbox-postgresql mode 'tcp'
+set load-balancing haproxy service sandbox-postgresql backend 'sandbox-postgresql-primary'
+set load-balancing haproxy backend sandbox-postgresql-primary mode 'tcp'
+set load-balancing haproxy backend sandbox-postgresql-primary health-check 'pgsql'
+set load-balancing haproxy backend sandbox-postgresql-primary server primary address {{ vyos_postgresql_primary_address }}
+set load-balancing haproxy backend sandbox-postgresql-primary server primary port {{ vyos_postgresql_port }}
+set load-balancing haproxy backend sandbox-postgresql-primary server primary check
+
 {# --- management --- #}
 set service ssh port 22
 

infrastructure/samba-ad/ansible/dns-forwarder.yml

@@ -0,0 +1,10 @@
+---
+- name: 更新现有 AD 域控 DNS 上游
+  hosts: samba_dc
+  become: true
+  gather_facts: false
+  tasks:
+    - name: 仅应用 DNS 转发器配置
+      ansible.builtin.import_role:
+        name: samba_ad_dc
+        tasks_from: dns_forwarder

infrastructure/samba-ad/ansible/group_vars/all/vars.yml

@@ -7,7 +7,7 @@ samba_ad_realm: "AD.DDUPAN.TOP"
 samba_ad_domain: "DDUPAN"
 samba_ad_dc_hostname: "dc1"
 samba_ad_dc_ip: "192.168.10.5"
-samba_ad_dns_forwarder: "192.168.10.1"
+samba_ad_dns_forwarder: "192.168.10.127"   # Blocky：DN42 条件转发及公网解析
 samba_ad_reverse_zone: "10.168.192.in-addr.arpa"   # reverse of 192.168.10.0/24
 
 # Static records now live in ../../dns/records.yml and are reconciled as complete

infrastructure/samba-ad/ansible/roles/samba_ad_dc/tasks/dns_forwarder.yml

@@ -0,0 +1,10 @@
+---
+- name: 更新 Samba DNS 转发器并校验配置
+  ansible.builtin.lineinfile:
+    path: /etc/samba/smb.conf
+    regexp: '^\s*dns forwarder\s*='
+    insertafter: '^\[global\]'
+    line: "\tdns forwarder = {{ samba_ad_dns_forwarder }}"
+    validate: '/usr/bin/testparm -s %s'
+    backup: true
+  notify: restart samba-ad-dc

infrastructure/samba-ad/ansible/roles/samba_ad_dc/tasks/main.yml

@@ -87,6 +87,9 @@
     creates: /var/lib/samba/private/sam.ldb
   no_log: true   # keep the admin password out of logs
 
+- name: 持续管理已初始化域控的 DNS 上游
+  ansible.builtin.import_tasks: dns_forwarder.yml
+
 - name: Install the generated krb5.conf system-wide
   ansible.builtin.copy:
     src: /var/lib/samba/private/krb5.conf