---
- name: Complete Nomad Cluster Reset and Rebuild
  hosts: nomad_cluster
  become: yes
  serial: 1  # 一次处理一个节点
  vars:
    nomad_encrypt_key: "NVOMDvXblgWfhtzFzOUIHnKEOrbXOkPrkIPbRGGf1YQ="
    tailscale_ips:
      semaphore: "100.116.158.95"
      master: "100.117.106.136" 
      ash3c: "100.116.80.94"
    
  tasks:
    - name: Stop nomad service completely
      systemd:
        name: nomad
        state: stopped
      ignore_errors: yes

    - name: Kill any remaining nomad processes
      shell: pkill -f nomad || true
      ignore_errors: yes

    - name: Remove all nomad data and state
      shell: |
        rm -rf /opt/nomad/data/*
        rm -rf /opt/nomad/data/.*
        rm -rf /var/log/nomad/*
      ignore_errors: yes

    - name: Create fresh nomad configuration with correct Tailscale IPs
      copy:
        content: |
          datacenter = "dc1"
          region     = "global"
          data_dir   = "/opt/nomad/data"
          
          # 使用 Tailscale IP 地址
          bind_addr = "{{ tailscale_ips[inventory_hostname] }}"
          
          server {
            enabled          = true
            bootstrap_expect = 3
            encrypt          = "{{ nomad_encrypt_key }}"
            
            server_join {
              retry_join = [
                "{{ tailscale_ips.semaphore }}",
                "{{ tailscale_ips.master }}",
                "{{ tailscale_ips.ash3c }}"
              ]
            }
          }
          
          client {
            enabled = true
            network_interface = "tailscale0"
          }
          
          ui_config {
            enabled = true
          }
          
          addresses {
            http = "0.0.0.0"
            rpc  = "{{ tailscale_ips[inventory_hostname] }}"
            serf = "{{ tailscale_ips[inventory_hostname] }}"
          }
          
          ports {
            http = 4646
            rpc  = 4647
            serf = 4648
          }
          
          plugin "docker" {
            config {
              allow_privileged = true
              volumes {
                enabled = true
              }
            }
          }
          
          log_level = "INFO"
          log_file  = "/var/log/nomad/nomad.log"
        dest: /etc/nomad.d/nomad.hcl
        owner: nomad
        group: nomad
        mode: '0640'

    - name: Ensure log directory exists
      file:
        path: /var/log/nomad
        state: directory
        owner: nomad
        group: nomad
        mode: '0755'

    - name: Start nomad service
      systemd:
        name: nomad
        state: started
        enabled: yes

    - name: Wait for nomad to start
      wait_for:
        port: 4646
        host: "{{ tailscale_ips[inventory_hostname] }}"
        delay: 5
        timeout: 30

    - name: Check nomad service status
      shell: systemctl status nomad --no-pager -l
      register: nomad_status
      ignore_errors: yes

    - name: Display nomad status
      debug:
        var: nomad_status.stdout_lines

- name: Wait for cluster to form
  hosts: localhost
  gather_facts: no
  tasks:
    - name: Wait for cluster formation
      pause:
        seconds: 30
        prompt: "等待集群形成..."

- name: Verify cluster status
  hosts: semaphore
  become: yes
  tasks:
    - name: Check cluster members
      shell: nomad server members
      register: cluster_members
      ignore_errors: yes

    - name: Display cluster members
      debug:
        var: cluster_members.stdout_lines

    - name: Check node status
      shell: nomad node status
      register: node_status
      ignore_errors: yes

    - name: Display node status
      debug:
        var: node_status.stdout_lines