Deploy CCM via Ansible before workers join to fix external cloud provider

This fixes the chicken-and-egg problem where workers with
--kubelet-arg=cloud-provider=external couldn't join because CCM wasn't
running yet to remove the node.cloudprovider.kubernetes.io/uninitialized taint.

Changes:
- Create ansible/roles/ccm-deploy/ to deploy CCM via Helm during Ansible phase
- Reorder site.yml: CCM deploys after secrets but before workers join
- CCM runs on control_plane[0] with proper tolerations for control plane nodes
- Add 10s pause after CCM ready to ensure it can process new nodes
- Workers can now successfully join with external cloud provider enabled

Flux still manages CCM for updates, but initial install happens in Ansible.

ansible/roles/ccm-deploy/tasks/main.yml

@@ -0,0 +1,62 @@
+---
+- name: Check if hcloud secret exists
+  command: kubectl -n kube-system get secret hcloud
+  register: hcloud_secret_check
+  changed_when: false
+  failed_when: false
+
+- name: Fail if hcloud secret is missing
+  fail:
+    msg: "hcloud secret not found in kube-system namespace. CCM requires it."
+  when: hcloud_secret_check.rc != 0
+
+- name: Add Hetzner Helm repository
+  kubernetes.core.helm_repository:
+    name: hcloud
+    repo_url: https://charts.hetzner.cloud
+    kubeconfig: /etc/rancher/k3s/k3s.yaml
+  environment:
+    KUBECONFIG: /etc/rancher/k3s/k3s.yaml
+
+- name: Deploy Hetzner Cloud Controller Manager
+  kubernetes.core.helm:
+    name: hcloud-cloud-controller-manager
+    chart_ref: hcloud/hcloud-cloud-controller-manager
+    release_namespace: kube-system
+    create_namespace: true
+    values:
+      networking:
+        enabled: true
+      nodeSelector:
+        kubernetes.io/hostname: "{{ inventory_hostname }}"
+      additionalTolerations:
+        - key: node-role.kubernetes.io/control-plane
+          operator: Exists
+          effect: NoSchedule
+    kubeconfig: /etc/rancher/k3s/k3s.yaml
+    wait: true
+    wait_timeout: 300s
+  environment:
+    KUBECONFIG: /etc/rancher/k3s/k3s.yaml
+
+- name: Wait for CCM to be ready
+  command: kubectl -n kube-system rollout status deployment/hcloud-cloud-controller-manager --timeout=120s
+  changed_when: false
+  register: ccm_rollout
+  until: ccm_rollout.rc == 0
+  retries: 3
+  delay: 10
+
+- name: Pause to ensure CCM is fully ready to process new nodes
+  pause:
+    seconds: 10
+
+- name: Verify CCM is removing uninitialized taints
+  command: kubectl get nodes -o jsonpath='{.items[*].spec.taints[?(@.key=="node.cloudprovider.kubernetes.io/uninitialized")].key}'
+  register: uninitialized_taints
+  changed_when: false
+  failed_when: false
+
+- name: Display taint status
+  debug:
+    msg: "Nodes with uninitialized taint: {{ uninitialized_taints.stdout }}"