first commit

cluster-experiments/docker-swarm/ansible/create-cluster.yml

@@ -0,0 +1,99 @@
+- name: Initialize Docker Swarm on the first manager
+  hosts: manager[0]
+  become: true
+  tasks:
+    - name: Check if node is already a manager
+      command: docker node inspect self --format '{{ "{{ .ManagerStatus }}" }}'
+      register: manager_status
+      ignore_errors: true
+      changed_when: false
+
+    - name: Initialize Docker Swarm if not already a manager
+      command: docker swarm init --advertise-addr {{ ansible_default_ipv4.address }}
+      register: swarm_init
+      when: manager_status.rc != 0
+      failed_when: swarm_init.rc != 0 and "'This node is already part of a swarm' not in swarm_init.stderr"
+      changed_when: "'Swarm initialized' in swarm_init.stdout"
+
+    - name: Get manager join-token
+      command: docker swarm join-token manager -q
+      register: manager_join_token
+
+    - name: Get worker join-token
+      command: docker swarm join-token worker -q
+      register: worker_join_token
+
+    - name: Set fact with join details
+      set_fact:
+        swarm_leader_ip: "{{ ansible_default_ipv4.address }}"
+        swarm_manager_token: "{{ manager_join_token.stdout }}"
+        swarm_worker_token: "{{ worker_join_token.stdout }}"
+
+    - name: Propagate join info globally
+      add_host:
+        name: "swarm_info_host"
+        swarm_leader_ip: "{{ swarm_leader_ip }}"
+        swarm_manager_token: "{{ swarm_manager_token }}"
+        swarm_worker_token: "{{ swarm_worker_token }}"
+      changed_when: false
+
+- name: Join additional managers to Swarm
+  hosts: swarm_managers:!manager[0]
+  become: true
+  vars:
+    swarm_info: "{{ hostvars['swarm_info_host'] }}"
+  tasks:
+    - name: Check if already in a swarm
+      command: docker info --format '{{ "{{ .Swarm.LocalNodeState }}" }}'
+      register: manager2_swarm_status
+      ignore_errors: true
+      changed_when: false
+
+    - name: Join as manager (if not already part of a swarm)
+      command: >
+        docker swarm join --token {{ swarm_info['swarm_manager_token'] }}
+        {{ swarm_info['swarm_leader_ip'] }}:2377
+      when: manager2_swarm_status.stdout != "active"
+
+- name: Join swarm_workers to Swarm
+  hosts: swarm_workers
+  become: true
+  vars:
+    swarm_info: "{{ hostvars['swarm_info_host'] }}"
+  tasks:
+    - name: Check if node is already part of a swarm
+      command: docker info --format '{{ "{{ .Swarm.LocalNodeState }}" }}'
+      register: swarm_worker_status
+      ignore_errors: true
+      changed_when: false
+
+    - name: Join as worker (if not already part of a swarm)
+      command: >
+        docker swarm join --token {{ swarm_info['swarm_worker_token'] }}
+        {{ swarm_info['swarm_leader_ip'] }}:2377
+      when: swarm_worker_status.stdout != "active"
+
+- name: Verify all nodes are present in the swarm
+  hosts: manager[0]
+  become: true
+  tasks:
+    - name: Get list of all nodes and their roles in the swarm
+      command: docker node ls --format '{{ "{{ .Hostname }}|{{ .ManagerStatus }}" }}'
+      register: swarm_nodes
+    - name: Parse manager and worker counts
+      set_fact:
+        manager_count: "{{ swarm_nodes.stdout_lines | select('search', 'Leader|Reachable') | list | length }}"
+        worker_count: "{{ swarm_nodes.stdout_lines | select('search', '^((?!Leader|Reachable).)*$') | list | length }}"
+    - name: Show parsed node roles
+      debug:
+        msg:
+          - "Managers found: {{ manager_count }} (expected: {{ groups['swarm_managers'] | length }})"
+          - "Workers found: {{ worker_count }} (expected: {{ groups['swarm_workers'] | length }})"
+    - name: Fail if manager or worker counts do not match inventory
+      fail:
+        msg: |
+          Swarm node count mismatch!
+          Managers found: {{ manager_count }} (expected: {{ groups['swarm_managers'] | length }})
+          Workers found: {{ worker_count }} (expected: {{ groups['swarm_workers'] | length }})
+          All nodes: {{ swarm_nodes.stdout_lines }}
+      when: (manager_count | int) != (groups['swarm_managers'] | length) or (worker_count | int) != (groups['swarm_workers'] | length)

cluster-experiments/docker-swarm/ansible/install-nfs-commons.yml

@@ -0,0 +1,16 @@
+- name: Ensure NFS client libraries are installed
+  hosts: all
+  become: yes
+
+  tasks:
+    - name: Install NFS client libraries (Debian/Ubuntu)
+      apt:
+        name: nfs-common
+        state: present
+      when: ansible_os_family == "Debian"
+
+    - name: Install NFS client libraries (RedHat/CentOS)
+      yum:
+        name: nfs-utils
+        state: present
+      when: ansible_os_family == "RedHat"

cluster-experiments/docker-swarm/ansible/update-cluster.yml

@@ -0,0 +1,86 @@
+- name: Gather join tokens from the primary manager
+  hosts: manager
+  become: true
+  tasks:
+    - name: Get manager join-token
+      command: docker swarm join-token manager -q
+      register: manager_join_token
+
+    - name: Get worker join-token
+      command: docker swarm join-token worker -q
+      register: worker_join_token
+
+    - name: Set fact with join details
+      set_fact:
+        swarm_leader_ip: "{{ ansible_default_ipv4.address }}"
+        swarm_manager_token: "{{ manager_join_token.stdout }}"
+        swarm_worker_token: "{{ worker_join_token.stdout }}"
+
+    - name: Propagate join info globally
+      add_host:
+        name: "swarm_info_host"
+        swarm_leader_ip: "{{ swarm_leader_ip }}"
+        swarm_manager_token: "{{ swarm_manager_token }}"
+        swarm_worker_token: "{{ swarm_worker_token }}"
+      changed_when: false
+
+- name: Add new managers to the Swarm
+  hosts: swarm_managers
+  become: true
+  vars:
+    swarm_info: "{{ hostvars['swarm_info_host'] }}"
+  tasks:
+    - name: Check if already in a swarm
+      command: docker info --format '{{ "{{ .Swarm.LocalNodeState }}" }}'
+      register: manager_swarm_status
+      ignore_errors: true
+      changed_when: false
+
+    - name: Join as manager (if not already part of a swarm)
+      command: >
+        docker swarm join --token {{ swarm_info['swarm_manager_token'] }}
+        {{ swarm_info['swarm_leader_ip'] }}:2377
+      when: manager_swarm_status.stdout != "active"
+
+- name: Add new workers to the Swarm
+  hosts: swarm_workers
+  become: true
+  vars:
+    swarm_info: "{{ hostvars['swarm_info_host'] }}"
+  tasks:
+    - name: Check if node is already part of a swarm
+      command: docker info --format '{{ "{{ .Swarm.LocalNodeState }}" }}'
+      register: worker_swarm_status
+      ignore_errors: true
+      changed_when: false
+
+    - name: Join as worker (if not already part of a swarm)
+      command: >
+        docker swarm join --token {{ swarm_info['swarm_worker_token'] }}
+        {{ swarm_info['swarm_leader_ip'] }}:2377
+      when: worker_swarm_status.stdout != "active"
+
+- name: Verify all nodes are present in the swarm
+  hosts: manager[0]
+  become: true
+  tasks:
+    - name: Get list of all nodes and their roles in the swarm
+      command: docker node ls --format '{{ "{{ .Hostname }}|{{ .ManagerStatus }}" }}'
+      register: swarm_nodes
+    - name: Parse manager and worker counts
+      set_fact:
+        manager_count: "{{ swarm_nodes.stdout_lines | select('search', 'Leader|Reachable') | list | length }}"
+        worker_count: "{{ swarm_nodes.stdout_lines | select('search', '^((?!Leader|Reachable).)*$') | list | length }}"
+    - name: Show parsed node roles
+      debug:
+        msg:
+          - "Managers found: {{ manager_count }} (expected: {{ groups['swarm_managers'] | length }})"
+          - "Workers found: {{ worker_count }} (expected: {{ groups['swarm_workers'] | length }})"
+    - name: Fail if manager or worker counts do not match inventory
+      fail:
+        msg: |
+          Swarm node count mismatch!
+          Managers found: {{ manager_count }} (expected: {{ groups['swarm_managers'] | length }})
+          Workers found: {{ worker_count }} (expected: {{ groups['swarm_workers'] | length }})
+          All nodes: {{ swarm_nodes.stdout_lines }}
+      when: (manager_count | int) != (groups['swarm_managers'] | length) or (worker_count | int) != (groups['swarm_workers'] | length)