Add HA option, change to yaml inventory, cleanup

- HA option for multiple server nodes using embedded etcd
- Switch to yaml inventory file for easier editing and combining vars
- Update to full ansible module names
- Change master/node names to server/agent
- Cleanup small linting errors
- Add reboot playbook which staggers reboot to keep HA cluster up
- Move playbooks to playbook directory

Signed-off-by: Derek Nola <derek.nola@suse.com>

roles/k3s/agent/tasks/main.yml

@@ -1,16 +1,15 @@
 ---
-
 - name: Copy K3s service file
   ansible.builtin.template:
     src: "k3s.service.j2"
-    dest: "{{ systemd_dir }}/k3s-node.service"
+    dest: "{{ systemd_dir }}/k3s-agent.service"
     owner: root
     group: root
     mode: 0755
 
 - name: Enable and check K3s service
   ansible.builtin.systemd:
-    name: k3s-node
+    name: k3s-agent
     daemon_reload: yes
     state: restarted
     enabled: yes

roles/k3s/agent/templates/k3s.service.j2

@@ -7,7 +7,7 @@ After=network-online.target
 Type=notify
 ExecStartPre=-/sbin/modprobe br_netfilter
 ExecStartPre=-/sbin/modprobe overlay
-ExecStart=/usr/local/bin/k3s agent --server https://{{ master_ip }}:6443 --token {{ hostvars[groups['master'][0]]['token'] }} {{ extra_agent_args | default("") }}
+ExecStart=/usr/local/bin/k3s agent --server https://{{ api_endpoint }}:{{ api_port }} --token {{ hostvars[groups['server'][0]]['token'] }} {{ extra_agent_args | default("") }}
 KillMode=process
 Delegate=yes
 # Having non-zero Limit*s causes performance problems due to accounting overhead

roles/k3s/master/defaults/main.yml

@@ -1,2 +0,0 @@
----
-k3s_server_location: /var/lib/rancher/k3s

roles/k3s/server/defaults/main.yml

@@ -0,0 +1,12 @@
+---
+k3s_server_location: /var/lib/rancher/k3s
+server_init_args: >-
+  {% if groups['server'] | length > 1 %}
+    --token {{ k3s_token }}
+    {% if ansible_host == hostvars[groups['server'][0]]['ansible_host'] | default(groups['server'][0]) %}
+      --cluster-init
+    {% else %}
+      --server https://{{ hostvars[groups['server'][0]]['ansible_host'] | default(groups['server'][0]) }}:{{ api_port }}
+    {% endif %}
+  {% endif %}
+  {{ extra_server_args | default('') }}

roles/k3s/server/tasks/main.yml

@@ -1,4 +1,25 @@
 ---
+- name: Init cluster using temporary service
+  ansible.builtin.command:
+    cmd: "systemd-run -p RestartSec=2 -p Restart=on-failure --unit=k3s-init k3s server {{ server_init_args }}"
+    creates: "{{ systemd_dir }}/k3s.service"
+
+- name: Verification
+  block:
+    - name: Verify that all nodes joined
+      ansible.builtin.command:
+        cmd: k3s kubectl get nodes -l "node-role.kubernetes.io/control-plane=true" -o=jsonpath="{.items[*].metadata.name}"
+      register: nodes
+      until: nodes.rc == 0 and (nodes.stdout.split() | length) == (groups['server'] | length)
+      retries: 20
+      delay: 10
+      changed_when: false
+  always:
+    - name: Kill the temporary init service
+      ansible.builtin.systemd:
+        name: k3s-init
+        state: stopped
+      failed_when: false
 
 - name: Copy K3s service file
   register: k3s_service
@@ -30,12 +51,12 @@
     path: "{{ k3s_server_location }}/server/node-token"
     mode: "g+rx,o+rx"
 
-- name: Read node-token from master
+- name: Read node-token from server
   ansible.builtin.slurp:
     path: "{{ k3s_server_location }}/server/node-token"
   register: node_token
 
-- name: Store Master node-token
+- name: Store server node-token
   ansible.builtin.set_fact:
     token: "{{ node_token.content | b64decode | regex_replace('\n', '') }}"
 
@@ -59,10 +80,10 @@
     owner: "{{ ansible_user }}"
     mode: "u=rw,g=,o="
 
-- name: Replace https://localhost:6443 by https://master-ip:6443
+- name: Change server to API endpoint instead of localhost
   ansible.builtin.command: >-
     /usr/local/bin/k3s kubectl config set-cluster default
-      --server=https://{{ master_ip }}:6443
+      --server=https://{{ api_endpoint }}:{{ api_port }}
       --kubeconfig ~{{ ansible_user }}/.kube/config
   changed_when: true