Crea una app de libro de registro con balanceo de cargas HTTP

En este ejemplo avanzado se muestra cómo compilar una app de libro de registro que usa Node.js para su frontend y MySQL para su backend. Además, la plantilla crea y conecta un balanceador de cargas HTTP que hace el balanceo de cargas en dos zonas y un escalador automático destinado a escalar automáticamente la aplicación.

Recursos de implementación con balanceo de cargas HTTP
Recursos de implementación con balanceo de cargas HTTP (haz clic para agrandar)

En este ejemplo se da por sentado que estás familiarizado con los contenedores de Docker, como con los recursos de Compute Engine, especialmente el balanceo de cargas HTTP, el ajuste de escala automático, los grupos de instancias administrados y las plantillas de instancias.

Para obtener más instructivos de introducción, consulta la Guía de introducción o la Guía paso a paso.

Antes de comenzar

Crea tus plantillas

En este ejemplo se inicia una implementación con diferentes tipos de recursos. Para comenzar, crearás plantillas que definan estos recursos por separado. Luego, utilizarás estas plantillas en tu configuración final.

Al final de este ejemplo, tendrás una implementación que contiene los siguientes recursos:

  • Una instancia única de Compute Engine para la máquina virtual MySQL del backend
  • Una plantilla de instancias que utiliza una imagen de Docker
  • Dos grupos de instancias administrados de escala automática en dos zonas diferentes, que ejecutan el servicio node.js frontend
  • Otro grupo de dos instancias administrado de escala automática que entrega datos estáticos
  • Una verificación de estado y un balanceador de cargas HTTP que distribuye tráfico entre los grupos de instancias administrados respectivos

Crea las plantillas de backend

El backend de esta aplicación es una instancia única de Compute Engine que ejecuta un contenedor MySQL Docker. Crea una plantilla que defina una instancia de Compute Engine que use una imagen optimizada para el contenedor. Asígnale un nombre al archivo container_vm.[py|jinja]:

Jinja



{% from 'container_helper.jinja' import GenerateManifest %}
{% set COMPUTE_URL_BASE = 'https://www.googleapis.com/compute/v1/' %}

resources:
- name: {{ env['name'] }}
  type: compute.v1.instance
  properties:
    zone: {{ properties['zone'] }}
    machineType: {{ COMPUTE_URL_BASE }}projects/{{ env['project'] }}/zones/{{ properties['zone'] }}/machineTypes/f1-micro
    metadata:
      items:
      - key: gce-container-declaration
        value: |
          {{ GenerateManifest(env['name'], properties['port'], properties['dockerImage'], properties['dockerEnv'])|indent(10) }}
    disks:
    - deviceName: boot
      type: PERSISTENT
      autoDelete: true
      boot: true
      initializeParams:
        diskName: {{ env['name'] }}-disk
        sourceImage: {{ COMPUTE_URL_BASE }}projects/cos-cloud/global/images/{{ properties['containerImage'] }}
    networkInterfaces:
    - accessConfigs:
      - name: external-nat
        type: ONE_TO_ONE_NAT
      network: {{ COMPUTE_URL_BASE }}projects/{{ env['project'] }}/global/networks/default
    serviceAccounts:
      - email: default
        scopes:
        - https://www.googleapis.com/auth/logging.write
        - https://www.googleapis.com/auth/monitoring.write

Python

# Copyright 2016 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

"""Creates a Container VM with the provided Container manifest."""

from container_helper import GenerateManifest

COMPUTE_URL_BASE = 'https://www.googleapis.com/compute/v1/'

def GlobalComputeUrl(project, collection, name):
  return ''.join([COMPUTE_URL_BASE, 'projects/', project,
                  '/global/', collection, '/', name])

def ZonalComputeUrl(project, zone, collection, name):
  return ''.join([COMPUTE_URL_BASE, 'projects/', project,
                  '/zones/', zone, '/', collection, '/', name])

def GenerateConfig(context):
  """Generate configuration."""

  base_name = context.env['name']

  # Properties for the container-based instance.
  instance = {
      'zone': context.properties['zone'],
      'machineType': ZonalComputeUrl(context.env['project'],
                                     context.properties['zone'],
                                     'machineTypes',
                                     'f1-micro'),
      'metadata': {
          'items': [{
              'key': 'gce-container-declaration',
              'value': GenerateManifest(context)
              }]
          },
      'disks': [{
          'deviceName': 'boot',
          'type': 'PERSISTENT',
          'autoDelete': True,
          'boot': True,
          'initializeParams': {
              'diskName': base_name + '-disk',
              'sourceImage': GlobalComputeUrl('cos-cloud',
                                              'images',
                                              context.properties[
                                                  'containerImage'])
              },
          }],
      'networkInterfaces': [{
          'accessConfigs': [{
              'name': 'external-nat',
              'type': 'ONE_TO_ONE_NAT'
              }],
          'network': GlobalComputeUrl(context.env['project'],
                                      'networks',
                                      'default')
          }],
        'serviceAccounts': [{
            'email': 'default',
            'scopes': ['https://www.googleapis.com/auth/logging.write']
            }]
      }

  # Resources to return.
  resources = {
      'resources': [{
          'name': base_name,
          'type': 'compute.v1.instance',
          'properties': instance
          }]
      }

  return resources

La plantilla define la cantidad de variables, como containerImage y manifest, que se completarán cuando definas tu configuración. Esta plantilla por sí sola simplemente crea una única instancia de máquina virtual.

Cuando usas imágenes de contenedor en instancias de Compute Engine, también debes proporcionar un archivo de manifiesto (diferente del manifiesto de Deployment Manager) a fin de describir a Compute Engine qué imagen de contenedor debe usar. Crea un método auxiliar llamado container_helper.[py|jinja] para definir de forma dinámica el manifiesto del contenedor:

Jinja



{% macro GenerateManifest(name, port, dockerImage, dockerEnv) -%}
apiVersion: v1
kind: Pod
metadata:
  name: {{ name }}
spec:
  containers:
  - name: {{ name }}
    image: {{ dockerImage }}
    ports:
    - hostPort: {{ port }}
      containerPort: {{ port }}
    {% if dockerEnv -%}
    env:
    {% for key, value in dockerEnv.items() -%}
    - name: {{ key }}
      value: '{{ value }}'
    {% endfor -%}
    {% endif -%}
{%- endmacro -%}

Python

# Copyright 2016 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

"""Helper methods for working with containers in config."""

import six
import yaml

def GenerateManifest(context):
  """Generates a Container Manifest given a Template context.

  Args:
    context: Template context, which must contain dockerImage and port
        properties, and an optional dockerEnv property.

  Returns:
    A Container Manifest as a YAML string.
  """
  env_list = []
  if 'dockerEnv' in context.properties:
    for key, value in six.iteritems(context.properties['dockerEnv']):
      env_list.append({'name': key, 'value': str(value)})

  manifest = {
      'apiVersion': 'v1',
      'kind': 'Pod',
      'metadata': {
          'name': str(context.env['name'])
          },
      'spec': {
          'containers': [{
              'name': str(context.env['name']),
              'image': context.properties['dockerImage'],
              'ports': [{
                  'hostPort': context.properties['port'],
                  'containerPort': context.properties['port']
                  }],
              }]
          }
      }

  if env_list:
    manifest['spec']['containers'][0]['env'] = env_list

  return yaml.dump(manifest, default_flow_style=False)

Crea las plantillas frontend

El frontend de esta app ejecuta Node.js y permite que los usuarios publiquen mensajes en la página web. Habrá dos grupos de instancias administrados que contendrán dos instancias cada uno: un grupo de instancias administradas principal y un grupo de instancias administradas secundario para el balanceo de cargas.

Para crear estas plantillas de frontend, sigue las siguientes instrucciones.

  1. Crea una plantilla de instancias.

    Necesitas un recurso de plantilla de instancias para crear un grupo de instancias administrado, es decir, un grupo de instancias de VM idénticas, administradas centralmente. En este ejemplo se crea un grupo de instancias administrado para las instancias node.js frontend, pero, primero, debes crear la plantilla de la instancia.

    Define un archivo llamado container_instance_template.[py|jinja]:

    Jinja

    
    
    {% from 'container_helper.jinja' import GenerateManifest %}
    {% set IT_NAME = env['name'] + '-it' %}
    
    resources:
    - name: {{ IT_NAME }}
      type: compute.v1.instanceTemplate
      properties:
        properties:
          metadata:
            items:
            - key: gce-container-declaration
              value: |
                {{ GenerateManifest(env['name'], properties['port'],properties['dockerImage'], properties['dockerEnv'])|indent(12) }}
          machineType: f1-micro
          disks:
          - deviceName: boot
            boot: true
            autoDelete: true
            mode: READ_WRITE
            type: PERSISTENT
            initializeParams:
              sourceImage: https://www.googleapis.com/compute/v1/projects/cos-cloud/global/images/{{ properties['containerImage'] }}
          networkInterfaces:
          - accessConfigs:
            - name: external-nat
              type: ONE_TO_ONE_NAT
            network: https://www.googleapis.com/compute/v1/projects/{{ env['project'] }}/global/networks/default
          serviceAccounts:
            - email: default
              scopes:
              - https://www.googleapis.com/auth/logging.write
              - https://www.googleapis.com/auth/monitoring.write
    outputs:
    - name: instanceTemplateSelfLink
      value: $(ref.{{ IT_NAME }}.selfLink)
    

    Python

    # Copyright 2016 Google Inc. All rights reserved.
    #
    # Licensed under the Apache License, Version 2.0 (the "License");
    # you may not use this file except in compliance with the License.
    # You may obtain a copy of the License at
    #
    #     http://www.apache.org/licenses/LICENSE-2.0
    #
    # Unless required by applicable law or agreed to in writing, software
    # distributed under the License is distributed on an "AS IS" BASIS,
    # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    # See the License for the specific language governing permissions and
    # limitations under the License.
    
    """Creates a Container VM with the provided Container manifest."""
    
    from container_helper import GenerateManifest
    
    def GenerateConfig(context):
      """Generates configuration."""
    
      image = ''.join(['https://www.googleapis.com/compute/v1/',
                       'projects/cos-cloud/global/images/',
                       context.properties['containerImage']])
      default_network = ''.join(['https://www.googleapis.com/compute/v1/projects/',
                                 context.env['project'],
                                 '/global/networks/default'])
    
      instance_template = {
          'name': context.env['name'] + '-it',
          'type': 'compute.v1.instanceTemplate',
          'properties': {
              'properties': {
                  'metadata': {
                      'items': [{
                          'key': 'gce-container-declaration',
                          'value': GenerateManifest(context)
                          }]
                      },
                  'machineType': 'f1-micro',
                  'disks': [{
                      'deviceName': 'boot',
                      'boot': True,
                      'autoDelete': True,
                      'mode': 'READ_WRITE',
                      'type': 'PERSISTENT',
                      'initializeParams': {'sourceImage': image}
                      }],
                  'networkInterfaces': [{
                      'accessConfigs': [{
                          'name': 'external-nat',
                          'type': 'ONE_TO_ONE_NAT'
                          }],
                      'network': default_network
                      }],
                    'serviceAccounts': [{
                        'email': 'default',
                        'scopes': ['https://www.googleapis.com/auth/logging.write']
                        }]
                  }
              }
          }
    
      outputs = [{'name': 'instanceTemplateSelfLink',
                  'value': '$(ref.' + instance_template['name'] + '.selfLink)'}]
    
      return {'resources': [instance_template], 'outputs': outputs}
    

  2. Crea un grupo de instancias administrado con ajuste de escala automático.

    Ahora que ya tienes una plantilla de instancias, puedes definir una plantilla que utilice la plantilla de instancia a fin de crear un grupo de instancias administrado y con escala automática. Crea un nuevo archivo llamado autoscaled_group.[py|jinja] con los siguientes contenidos:

    Jinja

    
    
    resources:
      - name: {{ env["name"] }}-igm
        type: compute.v1.instanceGroupManager
        properties:
          zone: {{ properties["zone"] }}
          targetSize: {{ properties["size"] }}
          baseInstanceName: {{ env["name"] }}-instance
          instanceTemplate: {{ properties["instanceTemplate"] }}
    
      - name: {{ env["name"] }}-as
        type: compute.v1.autoscaler
        properties:
          zone: {{ properties["zone"] }}
          target: $(ref.{{ env["name"] }}-igm.selfLink)
          autoscalingPolicy:
            maxNumReplicas: {{ properties["maxSize"] }}
    

    Python

    # Copyright 2016 Google Inc. All rights reserved.
    #
    # Licensed under the Apache License, Version 2.0 (the "License");
    # you may not use this file except in compliance with the License.
    # You may obtain a copy of the License at
    #
    #     http://www.apache.org/licenses/LICENSE-2.0
    #
    # Unless required by applicable law or agreed to in writing, software
    # distributed under the License is distributed on an "AS IS" BASIS,
    # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    # See the License for the specific language governing permissions and
    # limitations under the License.
    
    """Creates autoscaled, network LB IGM running specified docker image."""
    
    def GenerateConfig(context):
      """Generate YAML resource configuration."""
    
      # NOTE: Once we can specify the port/service during creation of IGM,
      # we will wire it up here.
      name = context.env['name']
      resources = [{
          'name': name + '-igm',
          'type': 'compute.v1.instanceGroupManager',
          'properties': {
              'zone': context.properties['zone'],
              'targetSize': context.properties['size'],
              'baseInstanceName': name + '-instance',
              'instanceTemplate': context.properties['instanceTemplate']
          }
      }, {
          'name': name + '-as',
          'type': 'compute.v1.autoscaler',
          'properties': {
              'zone': context.properties['zone'],
              'target': '$(ref.' + name + '-igm.selfLink)',
              'autoscalingPolicy': {
                  'maxNumReplicas': context.properties['maxSize']
    
              }
          }
      }]
      return {'resources': resources}
    

    Crea el archivo de esquema correspondiente:

    Jinja

    # Copyright 2016 Google Inc. All rights reserved.
    #
    # Licensed under the Apache License, Version 2.0 (the "License");
    # you may not use this file except in compliance with the License.
    # You may obtain a copy of the License at
    #
    #     http://www.apache.org/licenses/LICENSE-2.0
    #
    # Unless required by applicable law or agreed to in writing, software
    # distributed under the License is distributed on an "AS IS" BASIS,
    # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    # See the License for the specific language governing permissions and
    # limitations under the License.
    
    info:
      title: Autoscaled, network LB IGM template
      author: Google
      description: Creates an autoscaled Instance Group Manager running the specified Docker image
      version: 1.0
    
    required:
    - zone
    - instanceTemplate
    
    properties:
      zone:
        type: string
        description: Zone in which this VM will run
    
      instanceTemplate:
        type: string
        description: URL for the instance template to use for IGM
    
      size:
        type: integer
        default: 1
        description: Initial size of the Managed Instance Group
    
      maxSize:
        type: integer
        default: 1
        description: Maximum size the Managed Instance Group will be autoscaled to
    

    Python

    # Copyright 2016 Google Inc. All rights reserved.
    #
    # Licensed under the Apache License, Version 2.0 (the "License");
    # you may not use this file except in compliance with the License.
    # You may obtain a copy of the License at
    #
    #     http://www.apache.org/licenses/LICENSE-2.0
    #
    # Unless required by applicable law or agreed to in writing, software
    # distributed under the License is distributed on an "AS IS" BASIS,
    # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    # See the License for the specific language governing permissions and
    # limitations under the License.
    
    info:
      title: Autoscaled, network LB IGM template
      author: Google
      description: Creates an autoscaled Instance Group Manager running the specified Docker image
      version: 1.0
    
    required:
    - zone
    - instanceTemplate
    
    properties:
      zone:
        type: string
        description: Zone in which this VM will run
    
      instanceTemplate:
        type: string
        description: URL for the instance template to use for IGM
    
      size:
        type: integer
        default: 1
        description: Initial size of the Managed Instance Group
    
      maxSize:
        type: integer
        default: 1
        description: Maximum size the Managed Instance Group will be autoscaled to
    

  3. Crea recursos mediante estas plantillas.

    Hasta este momento, definiste plantillas base que determinan las propiedades de tus recursos. Mediante estas plantillas, define la configuración de tu frontend. Crea un nuevo archivo llamado service.[py|jinja] con los siguientes contenidos:

    Jinja

    
    
    resources:
      - name: {{ env["name"] }}
        type: container_instance_template.jinja
        properties:
          port: {{ properties["port"] }}
          dockerEnv: {{ properties["dockerEnv"] }}
          dockerImage: {{ properties["dockerImage"] }}
          containerImage: {{ properties["containerImage"] }}
    
      - name: {{ env["name"] }}-pri
        type: autoscaled_group.jinja
        properties:
          zone: {{ properties["primaryZone"] }}
          size: {{ properties["primarySize"] }}
          maxSize: {{ properties["maxSize"] }}
          port: {{ properties["port"] }}
          service: {{ properties["service"] }}
          baseInstanceName: {{ env["name"] }}-instance
          instanceTemplate: $(ref.{{ env["name"] }}-it.selfLink)
    
      - name: {{ env["name"] }}-sec
        type: autoscaled_group.jinja
        properties:
          zone: {{ properties["secondaryZone"] }}
          size: {{ properties["secondarySize"] }}
          maxSize: {{ properties["maxSize"] }}
          port: {{ properties["port"] }}
          service: {{ properties["service"] }}
          baseInstanceName: {{ env["name"] }}-instance
          instanceTemplate: $(ref.{{ env["name"] }}-it.selfLink)
    
      - name: {{ env["name"] }}-hc
        type: compute.v1.httpHealthCheck
        properties:
          port: {{ properties["port"] }}
          requestPath: /_ah/health
    
      - name: {{ env["name"] }}-bes
        type: compute.v1.backendService
        properties:
          port: {{ properties["port"] }}
          portName: {{ properties["service"] }}
          backends:
            - name: {{ env["name"] }}-primary
              group: $(ref.{{ env["name"] }}-pri-igm.instanceGroup)
            - name: {{ env["name"] }}-secondary
              group: $(ref.{{ env["name"] }}-sec-igm.instanceGroup)
          healthChecks: [ $(ref.{{ env["name"] }}-hc.selfLink) ]
    

    Python

    # Copyright 2016 Google Inc. All rights reserved.
    #
    # Licensed under the Apache License, Version 2.0 (the "License");
    # you may not use this file except in compliance with the License.
    # You may obtain a copy of the License at
    #
    #     http://www.apache.org/licenses/LICENSE-2.0
    #
    # Unless required by applicable law or agreed to in writing, software
    # distributed under the License is distributed on an "AS IS" BASIS,
    # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    # See the License for the specific language governing permissions and
    # limitations under the License.
    
    """Creates primary/secondary zone autoscaled IGM running specified container."""
    
    def GenerateConfig(context):
      """Generate YAML resource configuration."""
    
      name = context.env['name']
    
      resources = [{
          'name': name,
          'type': 'container_instance_template.py',
          'properties': {
              'port': context.properties['port'],
              'dockerEnv': context.properties['dockerEnv'],
              'dockerImage': context.properties['dockerImage'],
              'containerImage': context.properties['containerImage']
          }
      }, {
          'name': name + '-pri',
          'type': 'autoscaled_group.py',
          'properties': {
              'zone': context.properties['primaryZone'],
              'size': context.properties['primarySize'],
              'maxSize': context.properties['maxSize'],
              'port': context.properties['port'],
              'service': context.properties['service'],
              'baseInstanceName': name + '-instance',
              'instanceTemplate': '$(ref.' + name + '-it.selfLink)'
          }
      }, {
          'name': name + '-sec',
          'type': 'autoscaled_group.py',
          'properties': {
              'zone': context.properties['secondaryZone'],
              'size': context.properties['secondarySize'],
              'maxSize': context.properties['maxSize'],
              'port': context.properties['port'],
              'service': context.properties['service'],
              'baseInstanceName': name + '-instance',
              'instanceTemplate': '$(ref.' + name + '-it.selfLink)'
          }
      }, {
          'name': name + '-hc',
          'type': 'compute.v1.httpHealthCheck',
          'properties': {
              'port': context.properties['port'],
              'requestPath': '/_ah/health'
          }
      }, {
          'name': name + '-bes',
          'type': 'compute.v1.backendService',
          'properties': {
              'port': context.properties['port'],
              'portName': context.properties['service'],
              'backends': [{
                  'name': name + '-primary',
                  'group': '$(ref.' + name + '-pri-igm.instanceGroup)'
              }, {
                  'name': name + '-secondary',
                  'group': '$(ref.' + name + '-sec-igm.instanceGroup)'
              }],
              'healthChecks': ['$(ref.' + name + '-hc.selfLink)']
          }
      }]
      return {'resources': resources}
    

    Crea el archivo de esquema correspondiente:

    Jinja

    # Copyright 2016 Google Inc. All rights reserved.
    #
    # Licensed under the Apache License, Version 2.0 (the "License");
    # you may not use this file except in compliance with the License.
    # You may obtain a copy of the License at
    #
    #     http://www.apache.org/licenses/LICENSE-2.0
    #
    # Unless required by applicable law or agreed to in writing, software
    # distributed under the License is distributed on an "AS IS" BASIS,
    # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    # See the License for the specific language governing permissions and
    # limitations under the License.
    
    info:
      title: Autoscaled IGM
      author: Google
      description: Creates primary/secondary zone autoscaled IGM running specified container.
      version: 1.0
    
    imports:
    - path: autoscaled_group.jinja
    - path: ../../common/jinja/container_instance_template.jinja
      name: container_instance_template.jinja
    
    required:
    - port
    - service
    - primaryZone
    - secondaryZone
    - dockerImage
    
    properties:
      primarySize:
        type: integer
        default: 1
        description: The size of the primary autoscaled IGM
    
      secondarySize:
        type: integer
        default: 0
        description: The size of the secondary autoscaled IGM
    
      maxSize:
        type: integer
        default: 1
        description: The maximum size of the IGM
    
      containerImage:
        type: string
        default: family/cos-stable
        description: The container image to be used
    
      dockerEnv:
        type: object
        default: {}
        description: The container environment variables
    
      dockerImage:
        type: string
        description: the docker image to be used
    
      port:
        type: integer
        description: Port to expose on the container as well as on the load balancer (e.g., 8080)
    
      service:
        type: string
        description: Name of the service the port exposes for loadbalancing (backendService) purposes
    
      primaryZone:
        type: string
        description: Primary Zone in which to run the service
    
      secondaryZone:
        type: string
        description: Secondary Zone in which to run the service
    

    Python

    # Copyright 2016 Google Inc. All rights reserved.
    #
    # Licensed under the Apache License, Version 2.0 (the "License");
    # you may not use this file except in compliance with the License.
    # You may obtain a copy of the License at
    #
    #     http://www.apache.org/licenses/LICENSE-2.0
    #
    # Unless required by applicable law or agreed to in writing, software
    # distributed under the License is distributed on an "AS IS" BASIS,
    # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    # See the License for the specific language governing permissions and
    # limitations under the License.
    
    info:
      title: Autoscaled IGM
      author: Google
      description: Creates primary/secondary zone autoscaled IGM running specified container.
      version: 1.0
    
    imports:
    - path: autoscaled_group.py
    - path: ../../common/python/container_instance_template.py
      name: container_instance_template.py
    
    required:
    - port
    - service
    - primaryZone
    - secondaryZone
    - dockerImage
    
    properties:
      primarySize:
        type: integer
        default: 1
        description: The size of the primary autoscaled IGM
    
      secondarySize:
        type: integer
        default: 0
        description: The size of the secondary autoscaled IGM
    
      maxSize:
        type: integer
        default: 1
        description: The maximum size of the IGM
    
      containerImage:
        type: string
        default: family/cos-stable
        description: The container image to be used
    
      dockerEnv:
        type: object
        default: {}
        description: The container environment variables
    
      dockerImage:
        type: string
        description: the docker image to be used
    
      port:
        type: integer
        description: Port to expose on the container as well as on the load balancer (e.g., 8080)
    
      service:
        type: string
        description: Name of the service the port exposes for loadbalancing (backendService) purposes
    
      primaryZone:
        type: string
        description: Primary Zone in which to run the service
    
      secondaryZone:
        type: string
        description: Secondary Zone in which to run the service
    

    Veamos en detalle lo que crea esta plantilla:

    1. Dos grupos de instancias administrados, uno principal y otro secundario.

      En la plantilla, se usa la plantilla autoscaled_group.[py|jinja] para crear un grupo de instancias administrado de escala automática principales y secundarias.

    2. A continuación, la plantilla crea un servicio de backend y verificación de estado. Se requiere un servicio de backend para el balanceo de cargas HTTP, que define la capacidad de servicio de los grupos de instancias en ese servicio de backend. En este caso, los grupos de instancias administrados principales y secundarios son parte de este backend y se aplican las propiedades predeterminadas del servicio de backend.

      De forma predeterminada, un servicio de backend realiza el balanceo de cargas en función del uso de CPU por parte de los grupos de instancias relacionados, pero también puedes realizar el balanceo de cargas en función de la solicitud por segundo (RPS).

      Nota: Siempre es obligatorio contar con una verificación de estado cuando se crea un servicio de backend.

Crea una plantilla unificadora

Por último, crea una plantilla unificadora que combine las plantillas de backend y de frontend. Crea un archivo nuevo llamado application.[py|jinja]:

Jinja



{% set BACKEND = env["deployment"] + "-backend" %}
{% set FRONTEND = env["deployment"] + "-frontend" %}
{% set STATIC_SERVICE = env["deployment"] + "-static-service" %}
{% set APPLICATION = env["deployment"] + "-application" %}

{% set APPLICATION_PORT = 8080 %}
{% set LB_PORT = 8080 %}
{% set MYSQL_PORT = 8080 %}

{% set CONTAINER_IMAGE = "family/cos-stable" %}

resources:
- name: {{ BACKEND }}
  type: container_vm.jinja
  properties:
    zone: {{ properties["primaryZone"] }}
    dockerImage: {{ properties["backendImage"] }}
    containerImage: {{ CONTAINER_IMAGE }}
    port: {{ MYSQL_PORT }}

- name: {{ FRONTEND }}
  type: service.jinja
  properties:
    primaryZone: {{ properties["primaryZone"] }}
    primarySize: 2
    secondaryZone: {{ properties["secondaryZone"] }}
    secondarySize: 0
    dockerImage: {{ properties["frontendImage"] }}
    containerImage: {{ CONTAINER_IMAGE }}
    port: {{ APPLICATION_PORT }}
    service: http
    # If left out will default to 1
    maxSize: 20

    # Define the variables that are exposed to container as env variables.
    dockerEnv:
      SEVEN_SERVICE_MYSQL_PORT: {{ MYSQL_PORT }}
      SEVEN_SERVICE_PROXY_HOST: $(ref.{{ BACKEND }}.networkInterfaces[0].networkIP)

- name: {{ STATIC_SERVICE }}
  type: service.jinja
  properties:
    primaryZone: {{ properties["primaryZone"] }}
    primarySize: 2
    secondaryZone: {{ properties["secondaryZone"] }}
    secondarySize: 0
    dockerImage: {{ properties["staticImage"] }}
    containerImage: {{ CONTAINER_IMAGE }}
    port: {{ APPLICATION_PORT }}
    service: httpstatic
    # If left out will default to 1
    maxSize: 20

- name: {{ APPLICATION }}-urlmap
  type: compute.v1.urlMap
  properties:
    defaultService: $(ref.{{ FRONTEND }}-bes.selfLink)
    hostRules:
      - hosts: ["*"]
        pathMatcher: pathmap
    pathMatchers:
      - name: pathmap
        defaultService: $(ref.{{ FRONTEND }}-bes.selfLink)
        pathRules:
          - paths: ["/static", "/static/*"]
            service: $(ref.{{ STATIC_SERVICE }}-bes.selfLink)
- name: {{ APPLICATION }}-targetproxy
  type: compute.v1.targetHttpProxy
  properties:
    urlMap: $(ref.{{ APPLICATION }}-urlmap.selfLink)
- name: {{ APPLICATION }}-l7lb
  type: compute.v1.globalForwardingRule
  properties:
    IPProtocol: TCP
    portRange: {{ LB_PORT }}
    target: $(ref.{{ APPLICATION }}-targetproxy.selfLink)
- name: {{ APPLICATION }}-fw
  type: compute.v1.firewall
  properties:
    allowed:
      - IPProtocol: TCP
        ports: [ {{ LB_PORT }} ]
    sourceRanges: [ 0.0.0.0/0 ]

Python


# Copyright 2016 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

"""Create appplication template with back-end and front-end templates."""

def GenerateConfig(context):
  """Generate configuration."""

  backend = context.env['deployment'] + '-backend'
  frontend = context.env['deployment'] + '-frontend'
  static_service = context.env['deployment'] + '-static-service'
  application = context.env['deployment'] + '-application'

  container_image = 'family/cos-stable'

  application_port = 8080
  lb_port = 8080
  mysql_port = 8080

  resources = [{
      'name': backend,
      'type': 'container_vm.py',
      'properties': {
          'zone': context.properties['primaryZone'],
          'dockerImage': context.properties['backendImage'],
          'containerImage': container_image,
          'port': mysql_port
      }
  }, {
      'name': frontend,
      'type': 'service.py',
      'properties': {
          'primaryZone': context.properties['primaryZone'],
          'primarySize': 2,
          'secondaryZone': context.properties['secondaryZone'],
          'secondarySize': 0,
          'dockerImage': context.properties['frontendImage'],
          'containerImage': container_image,
          'port': application_port,
          'service': 'http',
          # If left out will default to 1
          'maxSize': 20,
          # Define the variables that are exposed to container as env variables.
          'dockerEnv': {
              'SEVEN_SERVICE_MYSQL_PORT': mysql_port,
              'SEVEN_SERVICE_PROXY_HOST': '$(ref.' + backend
                                          + '.networkInterfaces[0].networkIP)'
          }
      }
  }, {
      'name': static_service,
      'type': 'service.py',
      'properties': {
          'primaryZone': context.properties['primaryZone'],
          'primarySize': 2,
          'secondaryZone': context.properties['secondaryZone'],
          'secondarySize': 0,
          'dockerImage': context.properties['staticImage'],
          'containerImage': container_image,
          'port': application_port,
          'service': 'httpstatic',
          # If left out will default to 1
          'maxSize': 20
      }
  }, {
      'name': application + '-urlmap',
      'type': 'compute.v1.urlMap',
      'properties': {
          'defaultService': '$(ref.' + frontend + '-bes.selfLink)',
          'hostRules': [{
              'hosts': ['*'],
              'pathMatcher': 'pathmap'
          }],
          'pathMatchers': [{
              'name': 'pathmap',
              'defaultService': '$(ref.' + frontend + '-bes.selfLink)',
              'pathRules': [{
                  'paths': ['/static', '/static/*'],
                  'service': '$(ref.' + static_service + '-bes.selfLink)'
              }]
          }]
      }
  }, {
      'name': application + '-targetproxy',
      'type': 'compute.v1.targetHttpProxy',
      'properties': {
          'urlMap': '$(ref.' + application + '-urlmap.selfLink)'
      }
  }, {
      'name': application + '-l7lb',
      'type': 'compute.v1.globalForwardingRule',
      'properties': {
          'IPProtocol': 'TCP',
          'portRange': lb_port,
          'target': '$(ref.' + application + '-targetproxy.selfLink)'
      }
  }, {
      'name': application + '-fw',
      'type': 'compute.v1.firewall',
      'properties': {
          'allowed': [{
              'IPProtocol': 'TCP',
              'ports': [lb_port]
          }],
          'sourceRanges': ['0.0.0.0/0']
      }
  }]
  return {'resources': resources}

Crea un archivo con el esquema correspondiente:

Jinja

# Copyright 2016 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

info:
  title: Application Template
  author: Google
  description: Create application template with back-end and front-end templates
  version: 1.0

imports:
- path: service.jinja
- path: ../../common/jinja/container_vm.jinja
  name: container_vm.jinja

required:
- primaryZone
- secondaryZone
- backendImage
- frontendImage
- staticImage

properties:
  primaryZone:
    type: string
    description: Primary Zone in which to run the service

  secondaryZone:
    type: string
    description: Secondary Zone in which to run the service

  backendImage:
    type: string
    description: Docker image to use in the backend

  frontendImage:
    type: string
    description: Docker image to use in the frontend service

  staticImage:
    type: string
    description: Docker image to use in the static service

Python

# Copyright 2016 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

info:
  title: Application Template
  author: Google
  description: Create application template with back-end and front-end templates
  version: 1.0

imports:
- path: service.py
- path: ../../common/python/container_vm.py
  name: container_vm.py

required:
- primaryZone
- secondaryZone
- backendImage
- frontendImage
- staticImage

properties:
  primaryZone:
    type: string
    description: Primary Zone in which to run the service

  secondaryZone:
    type: string
    description: Secondary Zone in which to run the service

  backendImage:
    type: string
    description: Docker image to use in the backend

  frontendImage:
    type: string
    description: Docker image to use in the frontend service

  staticImage:
    type: string
    description: Docker image to use in the static service

Además del frontend y backend, la plantilla también define algunos recursos adicionales:

  1. Un servicio estático, con grupos de instancias administrados principales y secundarios. Este servicio estático entrega a una página web ubicada en la ruta de acceso /static en tu aplicación.

  2. Un recurso de mapa de URL. El balanceo de cargas HTTP requiere un mapa de URL a fin de asignar las diferentes URL con las rutas correctas. En este caso, la ruta de acceso predeterminada, indicada mediante la propiedad defaultService, es el servicio de backend que creaste antes. Si un usuario navega hasta /static, el mapa de la URL asignará la ruta de acceso con el servicio estático, como se define en la sección pathMatchers.

  3. Una regla de reenvío global y un Proxy HTTP de destino. Dado que a la aplicación se le están balanceando las cargas en dos zonas separadas, necesitarás una regla de reenvío global que entregue a una sola dirección IP externa. Además, se requiere un Proxy HTTP de destino para la configuración del balanceo de cargas HTTP.

  4. Una regla de firewall que permite el tráfico a través del puerto 8080.

Crea tu configuración

Ahora que tus plantillas y tus esquemas relacionados están listos, puedes crear una configuración que implemente estos recursos. Crea un archivo de configuración llamado application.yaml con los siguientes contenidos y reemplaza ZONE_TO_RUN y SECONDARY_ZONE_TO_RUN con las zonas principal y secundaria que elijas.

Jinja

# Copyright 2016 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

# Launches an autoscaled, load-balanced frontend in two zones running nodejs
# for serving traffic using L7 loadbalancing. Also launches a single MySQL
# container instance, wires the two together using references, and passes
# them as env variables to the underlying frontend Docker containers.
#
# NOTE: Due to the fact that IGM does not allow specifying service/port to
# created IG, you must run the following commands after creation of the
# template:
#
# export DEPLOYMENT=<DEPLOYMENT NAME>
# export PRIMARY_ZONE=<PRIMARY ZONE>
# export SECONDARY_ZONE=<SECONDARY ZONE>
#
# gcloud compute instance-groups unmanaged set-named-ports ${DEPLOYMENT}-frontend-pri-igm \
#  --named-ports http:8080,httpstatic:8080 \
#  --zone ${PRIMARY_ZONE}
#
# gcloud compute instance-groups unmanaged set-named-ports ${DEPLOYMENT}-frontend-sec-igm \
#  --named-ports http:8080,httpstatic:8080 \
#  --zone ${SECONDARY_ZONE}
#
# Then to see the IP that exposes the application, you can do:
# gcloud compute forwarding-rules list | grep application-${DEPLOYMENT}-l7lb

imports:
- path: application.jinja

resources:
- name: nodejs
  type: application.jinja
  properties:
    primaryZone: ZONE_TO_RUN
    secondaryZone: SECOND_ZONE_TO_RUN
    backendImage: gcr.io/deployment-manager-examples/mysql
    frontendImage: gcr.io/deployment-manager-examples/nodejsservice
    staticImage: gcr.io/deployment-manager-examples/nodejsservicestatic

Python

# Copyright 2016 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

# Launches an autoscaled, load-balanced frontend in two zones running nodejs
# for serving traffic using L7 loadbalancing. Also launches a single MySQL
# container instance, wires the two together using references, and passes
# them as env variables to the underlying frontend Docker containers.
#
# NOTE: Due to the fact that IGM does not allow specifying service/port to
# created IG, you must run the following commands after creation of the
# template:
#
# export DEPLOYMENT=<DEPLOYMENT NAME>
# export PRIMARY_ZONE=<PRIMARY ZONE>
# export SECONDARY_ZONE=<SECONDARY ZONE>
#
# gcloud compute instance-groups unmanaged set-named-ports ${DEPLOYMENT}-frontend-pri-igm \
#  --named-ports http:8080,httpstatic:8080 \
#  --zone ${PRIMARY_ZONE}
#
# gcloud compute instance-groups unmanaged set-named-ports ${DEPLOYMENT}-frontend-sec-igm \
#  --named-ports http:8080,httpstatic:8080 \
#  --zone ${SECONDARY_ZONE}
#
# Then to see the IP that exposes the application, you can do:
# gcloud compute forwarding-rules list | grep application-${DEPLOYMENT}-l7lb

imports:
- path: application.py

resources:
- name: nodejs
  type: application.py
  properties:
    primaryZone: ZONE_TO_RUN
    secondaryZone: SECOND_ZONE_TO_RUN
    backendImage: gcr.io/deployment-manager-examples/mysql
    frontendImage: gcr.io/deployment-manager-examples/nodejsservice
    staticImage: gcr.io/deployment-manager-examples/nodejsservicestatic

Implementando la configuración

Implementemos tus recursos. Con Google Cloud CLI, ejecuta el siguiente comando y, de manera opcional, reemplaza advanced-configuration-l7 por el nombre de implementación que elijas. Ten en cuenta que el nombre de la implementación se usará automáticamente para nombrar los recursos.

En este ejemplo, el nombre de implementación es advanced-configuration-l7. Si decides cambiar el nombre de la implementación, asegúrate de usar ese nombre en todos los siguientes ejemplos.

gcloud deployment-manager deployments create advanced-configuration-l7 --config application.yaml

La respuesta debería ser similar a los siguientes recursos:

Waiting for create operation-1469468950934-5387966d431f0-49b11bc4-1421b2f0...done.
Create operation operation-1469468950934-5387966d431f0-49b11bc4-1421b2f0 completed successfully.
NAME                                               TYPE                             STATE      ERRORS
advanced-configuration-l7-application-fw           compute.v1.firewall              COMPLETED  []
advanced-configuration-l7-application-l7lb         compute.v1.globalForwardingRule  COMPLETED  []
advanced-configuration-l7-application-targetproxy  compute.v1.targetHttpProxy       COMPLETED  []
advanced-configuration-l7-application-urlmap       compute.v1.urlMap                COMPLETED  []
advanced-configuration-l7-backend                  compute.v1.instance              COMPLETED  []
advanced-configuration-l7-frontend-bes             compute.v1.backendService        COMPLETED  []
advanced-configuration-l7-frontend-hc              compute.v1.httpHealthCheck       COMPLETED  []
advanced-configuration-l7-frontend-it              compute.v1.instanceTemplate      COMPLETED  []
advanced-configuration-l7-frontend-pri-as          compute.v1.autoscaler            COMPLETED  []
advanced-configuration-l7-frontend-pri-igm         compute.v1.instanceGroupManager  COMPLETED  []
advanced-configuration-l7-frontend-sec-as          compute.v1.autoscaler            COMPLETED  []
advanced-configuration-l7-frontend-sec-igm         compute.v1.instanceGroupManager  COMPLETED  []
advanced-configuration-l7-static-service-bes       compute.v1.backendService        COMPLETED  []
advanced-configuration-l7-static-service-hc        compute.v1.httpHealthCheck       COMPLETED  []
advanced-configuration-l7-static-service-it        compute.v1.instanceTemplate      COMPLETED  []
advanced-configuration-l7-static-service-pri-as    compute.v1.autoscaler            COMPLETED  []
advanced-configuration-l7-static-service-pri-igm   compute.v1.instanceGroupManager  COMPLETED  []
advanced-configuration-l7-static-service-sec-as    compute.v1.autoscaler            COMPLETED  []
advanced-configuration-l7-static-service-sec-igm   compute.v1.instanceGroupManager  COMPLETED  []

Agrega etiquetas de servicio

A continuación, especifica las etiquetas de servicio adecuadas para tus grupos de instancias administrados. Las etiquetas de servicio son metadatos utilizados por el servicio de balanceo de cargas para agrupar recursos.

A fin de agregar etiquetas de servicio, ejecuta los siguientes comandos, haciendo coincidir las zonas principales y secundarias que seleccionaste en tu archivo de configuración de implementación:

gcloud compute instance-groups unmanaged set-named-ports advanced-configuration-l7-frontend-pri-igm \
  --named-ports http:8080,httpstatic:8080 \
  --zone [PRIMARY_ZONE]

gcloud compute instance-groups unmanaged set-named-ports advanced-configuration-l7-frontend-sec-igm \
  --named-ports http:8080,httpstatic:8080 \
  --zone [SECONDARY_ZONE]

Prueba la configuración

A fin de probar tu configuración, obtén la dirección IP externa que está enviando tráfico; para ello, consulta la regla de reenvío:

gcloud compute forwarding-rules list | grep advanced-configuration-l7-l7lb
advanced-configuration-l7-l7lb             107.178.249.126 TCP         advanced-configuration-l7-targetproxy

En este caso, la IP externa es 107.178.249.126.

En un navegador, visita la dirección IP externa en el puerto 8080. Por ejemplo, si tu IP externa es 107.178.249.126, la URL sería la siguiente:

http://107.178.249.126:8080

Deberías ver una página en blanco. A continuación, publica un mensaje en la página. Ve a la siguiente URL:

http://107.178.249.126:8080?msg=hello_world!

Verás una confirmación de que se agregó tu mensaje. Navega nuevamente a la URL principal; ahora, la página debería tener el siguiente mensaje:

hello_world!

Además, puedes visitar la página estática que creaste, o verificar el estado de tu aplicación; para ello, visita las siguientes URL:

# Static web page
http://107.178.249.126:8080/static

# Health check
http://107.178.249.126:8080/_ah/health

Felicitaciones, has implementado tu configuración correctamente.

Crea imágenes de Docker (opcional)

Docker te permite automatizar y ejecutar software dentro de contenedores. Los contenedores te permiten aislar diferentes servicios dentro de contenedores que pueden ejecutarse todos en una única instancia de Linux.

En este ejemplo se usaron algunas imágenes existentes de Docker, pero también puedes crear tus propias versiones de estas imágenes de Docker. Puedes encontrar las instrucciones para crear imágenes backend MySQL y las imágenes de frontend de Node.js en la sección Crea tus plantillas de recursos.

A fin de crear una imagen de Docker que entregue a la página web estática, sigue estos pasos:

  1. Crea una nueva instancia de VM con una imagen optimizada para el contenedor:

    gcloud compute instances create docker-playground \
      --image-family container-vm \
      --image-project google-containers \
      --zone us-central1-a \
      --machine-type f1-micro
    
  2. Conéctate a la instancia:

    gcloud compute ssh --zone us-central1-a docker-playground
    
  3. Crea un archivo llamado Dockerfile con los siguientes contenidos:

    FROM node:latest
    
    RUN mkdir /var/www/
    ADD service.js /var/www/service.js
    WORKDIR /var/www/
    RUN npm install mysql
    
    CMD ["node", "service.js"]
    
  4. Crea un archivo llamado service.js con el siguiente contenido:

    var http = require('http');
    var url = require('url');
    
    console.log('Started static node server')
    
    http.createServer(function (req, res) {
      reqUrl = url.parse(req.url, true);
    
      res.useChunkedEncodingByDefault = false;
      res.writeHead(200, {'Content-Type': 'text/html'});
    
      if (reqUrl.pathname == '/_ah/health') {
        res.end('ok');
      } else if (reqUrl.pathname == '/exit') {
        process.exit(-1)
      } else {
          res.end('static server');
      }
    }).listen(8080, '0.0.0.0');
    
    console.log('Static server running at http://127.0.0.1:8080/');
    
  5. Compila la imagen de Docker y reemplaza username con tu nombre de usuario de Docker Hub. Si no tienes un nombre de usuario de Docker Hub, crea uno antes de compilar la imagen de Docker.

    sudo docker build --no-cache -t username/nodejsservicestatic .
    
  6. Envía las imágenes al repositorio de Docker:

    sudo docker push username/nodejsservicestatic
    

Ahora tienes las imágenes de Docker necesarias para ejecutar node.js y MySQL. Puedes ver estas imágenes en el repositorio si buscas los nombres de las imágenes. Para probar las imágenes, puedes reemplazar todas las instancias de gcr.io/deployment-manager-examples/mysql y gcr.io/deployment-manager-examples/nodejsservice con tus imágenes respectivas.

Próximos pasos

Una vez que hayas completado este ejemplo, podrás hacer lo siguiente:

  • Continuar compilando este ejemplo; para ello, crea una página web más sólida o agrega más servicios al servidor web
  • Obtén más información acerca de las configuraciones o las implementaciones.
  • Intentar crear tus propias configuraciones