michaelschiemer/deployment/ansible/playbooks/setup/redeploy-traefik-gitea-clean.yml

---
# Clean Redeploy Traefik and Gitea Stacks
# Complete redeployment with backup, container recreation, and verification
#
# Usage:
#   # With automatic backup
#   ansible-playbook -i inventory/production.yml playbooks/setup/redeploy-traefik-gitea-clean.yml \
#     --vault-password-file secrets/.vault_pass
#
#   # With existing backup
#   ansible-playbook -i inventory/production.yml playbooks/setup/redeploy-traefik-gitea-clean.yml \
#     --vault-password-file secrets/.vault_pass \
#     -e "backup_name=redeploy-backup-1234567890" \
#     -e "skip_backup=true"

- name: Clean Redeploy Traefik and Gitea
  hosts: production
  gather_facts: yes
  become: no
  vars:
    traefik_stack_path: "{{ stacks_base_path }}/traefik"
    gitea_stack_path: "{{ stacks_base_path }}/gitea"
    gitea_url: "https://{{ gitea_domain }}"
    traefik_container_name: "traefik"
    gitea_container_name: "gitea"
    backup_base_path: "{{ backups_path | default('/home/deploy/backups') }}"
    skip_backup: "{{ skip_backup | default(false) | bool }}"
    backup_name: "{{ backup_name | default('') }}"

  tasks:
    # ========================================
    # 1. BACKUP (unless skipped)
    # ========================================
    - name: Set backup name fact
      ansible.builtin.set_fact:
        actual_backup_name: "{{ backup_name | default('redeploy-backup-' + ansible_date_time.epoch) }}"
      when: not skip_backup

    - name: Display backup note
      ansible.builtin.debug:
        msg: |
          ⚠️  NOTE: Backup should be run separately before redeploy:
          ansible-playbook -i inventory/production.yml playbooks/maintenance/backup-before-redeploy.yml \
            --vault-password-file secrets/.vault_pass \
            -e "backup_name={{ actual_backup_name }}"
          
          Or use existing backup with: -e "backup_name=redeploy-backup-XXXXX" -e "skip_backup=true"
      when: not skip_backup

    - name: Display redeployment plan
      ansible.builtin.debug:
        msg: |
          ================================================================================
          CLEAN REDEPLOY TRAEFIK AND GITEA
          ================================================================================
          
          This playbook will:
          1. ✅ Backup ({% if skip_backup %}SKIPPED{% else %}Performed{% endif %})
          2. ✅ Stop and remove Traefik containers (keeps acme.json)
          3. ✅ Stop and remove Gitea containers (keeps volumes/data)
          4. ✅ Sync latest stack configurations
          5. ✅ Redeploy Traefik stack
          6. ✅ Redeploy Gitea stack
          7. ✅ Restore Gitea configuration (app.ini)
          8. ✅ Verify service discovery
          9. ✅ Test Gitea accessibility
          
          ⚠️  IMPORTANT:
          - SSL certificates (acme.json) will be preserved
          - Gitea data (volumes) will be preserved
          - Only containers will be recreated
          - Expected downtime: ~2-5 minutes
          {% if not skip_backup %}
          - Backup location: {{ backup_base_path }}/{{ actual_backup_name }}
          {% endif %}
          
          ================================================================================

    # ========================================
    # 2. STOP AND REMOVE CONTAINERS
    # ========================================
    - name: Stop Traefik stack
      ansible.builtin.shell: |
        cd {{ traefik_stack_path }}
        docker compose down
      register: traefik_stop
      changed_when: traefik_stop.rc == 0
      failed_when: false

    - name: Remove Traefik containers (if any remain)
      ansible.builtin.shell: |
        docker ps -a --filter "name={{ traefik_container_name }}" --format "{{ '{{' }}.ID{{ '}}' }}" | xargs -r docker rm -f 2>/dev/null || true
      register: traefik_remove
      changed_when: traefik_remove.rc == 0
      failed_when: false

    - name: Stop Gitea stack (preserves volumes)
      ansible.builtin.shell: |
        cd {{ gitea_stack_path }}
        docker compose down
      register: gitea_stop
      changed_when: gitea_stop.rc == 0
      failed_when: false

    - name: Remove Gitea containers (if any remain, volumes are preserved)
      ansible.builtin.shell: |
        docker ps -a --filter "name={{ gitea_container_name }}" --format "{{ '{{' }}.ID{{ '}}' }}" | xargs -r docker rm -f 2>/dev/null || true
      register: gitea_remove
      changed_when: gitea_remove.rc == 0
      failed_when: false

    # ========================================
    # 3. SYNC CONFIGURATIONS
    # ========================================
    - name: Get stacks directory path
      ansible.builtin.set_fact:
        stacks_source_path: "{{ playbook_dir | dirname | dirname | dirname }}/stacks"
      delegate_to: localhost
      run_once: true

    - name: Sync stacks directory to production server
      ansible.builtin.synchronize:
        src: "{{ stacks_source_path }}/"
        dest: "{{ stacks_base_path }}/"
        delete: no
        recursive: yes
        rsync_opts:
          - "--chmod=D755,F644"
          - "--exclude=.git"
          - "--exclude=*.log"
          - "--exclude=data/"
          - "--exclude=volumes/"
          - "--exclude=acme.json"  # Preserve SSL certificates
          - "--exclude=*.key"
          - "--exclude=*.pem"

    # ========================================
    # 4. ENSURE ACME.JSON EXISTS
    # ========================================
    - name: Check if acme.json exists
      ansible.builtin.stat:
        path: "{{ traefik_stack_path }}/acme.json"
      register: acme_json_stat

    - name: Ensure acme.json exists and has correct permissions
      ansible.builtin.file:
        path: "{{ traefik_stack_path }}/acme.json"
        state: touch
        mode: '0600'
        owner: "{{ ansible_user }}"
        group: "{{ ansible_user }}"
      become: yes
      register: acme_json_ensure

    # ========================================
    # 5. REDEPLOY TRAEFIK
    # ========================================
    - name: Deploy Traefik stack
      community.docker.docker_compose_v2:
        project_src: "{{ traefik_stack_path }}"
        state: present
        pull: always
      register: traefik_deploy

    - name: Wait for Traefik to be ready
      ansible.builtin.shell: |
        cd {{ traefik_stack_path }}
        docker compose ps {{ traefik_container_name }} | grep -Eiq "Up|running"
      register: traefik_ready
      changed_when: false
      until: traefik_ready.rc == 0
      retries: 12
      delay: 5
      failed_when: traefik_ready.rc != 0

    # ========================================
    # 6. REDEPLOY GITEA
    # ========================================
    - name: Deploy Gitea stack
      community.docker.docker_compose_v2:
        project_src: "{{ gitea_stack_path }}"
        state: present
        pull: always
      register: gitea_deploy

    - name: Wait for Gitea to be ready
      ansible.builtin.shell: |
        cd {{ gitea_stack_path }}
        docker compose ps {{ gitea_container_name }} | grep -Eiq "Up|running"
      register: gitea_ready
      changed_when: false
      until: gitea_ready.rc == 0
      retries: 12
      delay: 5
      failed_when: gitea_ready.rc != 0

    - name: Wait for Gitea to be healthy
      ansible.builtin.shell: |
        cd {{ gitea_stack_path }}
        docker compose exec -T {{ gitea_container_name }} curl -f http://localhost:3000/api/healthz 2>&1 | grep -q "status.*pass" && echo "HEALTHY" || echo "NOT_HEALTHY"
      register: gitea_health
      changed_when: false
      until: gitea_health.stdout == "HEALTHY"
      retries: 30
      delay: 2
      failed_when: false

    # ========================================
    # 7. RESTORE GITEA CONFIGURATION
    # ========================================
    - name: Restore Gitea app.ini from backup
      ansible.builtin.shell: |
        if [ -f "{{ backup_base_path }}/{{ actual_backup_name }}/gitea-app.ini" ]; then
          cd {{ gitea_stack_path }}
          docker compose exec -T {{ gitea_container_name }} sh -c "cat > /data/gitea/conf/app.ini" < "{{ backup_base_path }}/{{ actual_backup_name }}/gitea-app.ini"
          docker compose restart {{ gitea_container_name }}
          echo "app.ini restored and Gitea restarted"
        else
          echo "No app.ini backup found, using default configuration"
        fi
      when: not skip_backup
      register: gitea_app_ini_restore
      changed_when: false
      failed_when: false

    # ========================================
    # 8. VERIFY SERVICE DISCOVERY
    # ========================================
    - name: Wait for service discovery (Traefik needs time to discover Gitea)
      ansible.builtin.pause:
        seconds: 15

    - name: Check if Gitea is in traefik-public network
      ansible.builtin.shell: |
        docker network inspect traefik-public --format '{{ '{{' }}range .Containers{{ '}}' }}{{ '{{' }}.Name{{ '}}' }} {{ '{{' }}end{{ '}}' }}' 2>/dev/null | grep -q {{ gitea_container_name }} && echo "YES" || echo "NO"
      register: gitea_in_network
      changed_when: false

    - name: Test direct connection from Traefik to Gitea
      ansible.builtin.shell: |
        cd {{ traefik_stack_path }}
        docker compose exec -T {{ traefik_container_name }} wget -qO- --timeout=5 http://{{ gitea_container_name }}:3000/api/healthz 2>&1 | head -5 || echo "CONNECTION_FAILED"
      register: traefik_gitea_direct
      changed_when: false
      failed_when: false

    # ========================================
    # 9. FINAL VERIFICATION
    # ========================================
    - name: Test Gitea via HTTPS (with retries)
      ansible.builtin.uri:
        url: "{{ gitea_url }}/api/healthz"
        method: GET
        status_code: [200]
        validate_certs: false
        timeout: 10
      register: gitea_https_test
      until: gitea_https_test.status == 200
      retries: 20
      delay: 3
      changed_when: false
      failed_when: false

    - name: Check SSL certificate status
      ansible.builtin.shell: |
        cd {{ traefik_stack_path }}
        if [ -f acme.json ] && [ -s acme.json ]; then
          echo "SSL certificates: PRESENT"
        else
          echo "SSL certificates: MISSING or EMPTY"
        fi
      register: ssl_status
      changed_when: false

    - name: Final status summary
      ansible.builtin.debug:
        msg: |
          ================================================================================
          REDEPLOYMENT SUMMARY
          ================================================================================
          
          Traefik:
          - Status: {{ traefik_ready.rc | ternary('Up', 'Down') }}
          - SSL Certificates: {{ ssl_status.stdout }}
          
          Gitea:
          - Status: {{ gitea_ready.rc | ternary('Up', 'Down') }}
          - Health: {% if gitea_health.stdout == 'HEALTHY' %}✅ Healthy{% else %}❌ Not Healthy{% endif %}
          - Configuration: {% if gitea_app_ini_restore.changed %}✅ Restored{% else %}ℹ️  Using default{% endif %}
          
          Service Discovery:
          - Gitea in network: {% if gitea_in_network.stdout == 'YES' %}✅{% else %}❌{% endif %}
          - Direct connection: {% if 'CONNECTION_FAILED' not in traefik_gitea_direct.stdout %}✅{% else %}❌{% endif %}
          
          Gitea Accessibility:
          {% if gitea_https_test.status == 200 %}
          ✅ Gitea is reachable via HTTPS (Status: 200)
          URL: {{ gitea_url }}
          {% else %}
          ❌ Gitea is NOT reachable via HTTPS (Status: {{ gitea_https_test.status | default('TIMEOUT') }})
          
          Possible causes:
          1. SSL certificate is still being generated (wait 2-5 minutes)
          2. Service discovery needs more time (wait 1-2 minutes)
          3. Network configuration issue
          
          Next steps:
          - Wait 2-5 minutes and test again: curl -k {{ gitea_url }}/api/healthz
          - Check Traefik logs: cd {{ traefik_stack_path }} && docker compose logs {{ traefik_container_name }} --tail=50
          - Check Gitea logs: cd {{ gitea_stack_path }} && docker compose logs {{ gitea_container_name }} --tail=50
          {% endif %}
          
          {% if not skip_backup %}
          Backup location: {{ backup_base_path }}/{{ actual_backup_name }}
          To rollback: ansible-playbook -i inventory/production.yml playbooks/maintenance/rollback-redeploy.yml \
            --vault-password-file secrets/.vault_pass \
            -e "backup_name={{ actual_backup_name }}"
          {% endif %}
          
          ================================================================================