Help & Documentation>Serverless Cloud Function>Web Framework Development>Deploying Framework On Command Line

Deploying Framework On Command Line

Last updated: 2023-09-27 14:52:26

On this page:

In addition to the console, you can also rapidly deploy a Web framework via the command line. This document will specifically guide you on how to locally deploy a Web application using the HTTP component of the Serverless Cloud Framework.

Preparations

You have installed the Serverless Cloud Framework, activated the service, and completed the permission configuration for the Serverless Cloud Framework.

Supported Framework

Supported Framework
Documentation
Express
Koa
Egg
Next.js
Nuxt.js
Nest.js
Flask
Django
Laravel

Instructions

1. Developing Applications Locally Based on your actual business scenario, complete the development locally. For more details, please refer to the Supported Frameworks development documentation.
2. Configuring the YML File In the project root directory, create a new serverless.yml file and write the configuration according to the following example. For comprehensive configuration, please refer to the Configuration Document.
# serverless.yml
component: http # Component name, which is required
name: webDemo # Component instance name, which is required.

inputs:
  region: ap-guangzhou # The region where the cloud function is located
  src: # Deploy the code files under src, package them into a zip file, and upload to the bucket.
    src: ./ # The local directory of files that need to be packaged
    exclude: # Files or directories to be excluded
      - .env
      - 'node_modules/**'
  faas: # Function configuration related
    framework: express # Select framework, using express as an example here
    runtime: Nodejs12.16
    name: webDemo # Cloud function name
    timeout: 10 # Timeout period, in seconds
    memorySize: 512 # Memory size, default is 512 MB
    layers:
      - name: layerName # Layer name
        version: 1 # Version

  apigw: # The HTTP component will automatically create an API Gateway service by default
    isDisabled: false # Whether to disable the automatic creation of API Gateway feature
    id: service-xxx # API Gateway service ID, if not filled, a new gateway will be automatically created
    name: serverless # API Gateway service ID
    api: # Configuration related to the created API
      cors: true # Allow cross-origin resource sharing (CORS)
      timeout: 15 # API timeout duration
      name: apiName # API name
      qualifier: $DEFAULT # Version associated with the API
    protocols:
      - http
      - https
    environment: test

3. Upon completion, execute scf deploy in the root directory for deployment. The component will automatically generate the scf_bootstrap startup file for deployment based on the selected framework type.
Note
As the bootstrap file logic is strongly related to your business logic, the generated default bootstrap file may cause the framework start to fail. We recommend you manually configure the bootstrap file according to your actual business needs. For more information, please see the deployment guide document of each framework.

Example scf_bootstrap

express:
#!/usr/bin/env bash

/var/lang/node12/bin/node app.js
koa
#!/usr/bin/env bash

/var/lang/node12/bin/node app.js
egg
#!/var/lang/node12/bin/node

/**
 * Node path in Docker: /var/lang/node12/bin/node
 * Since the serverless function only has read and write permissions for /tmp, it is necessary to modify two environment variables during startup.
 * NODE_LOG_DIR changes the default node write path of egg-scripts (~/logs) to /tmp
 * EGG_APP_CONFIG changes the default directory of the Egg application to /tmp
 */

process.env.EGG_SERVER_ENV = 'prod';
process.env.NODE_ENV = 'production';
process.env.NODE_LOG_DIR = '/tmp';
process.env.EGG_APP_CONFIG = '{"rundir":"/tmp","logger":{"dir":"/tmp"}}';

const { Application } = require('egg');

// If you deploy node_modules through layers, you need to modify eggPath
Object.defineProperty(Application.prototype, Symbol.for('egg#eggPath'), {
  value: '/opt',
});

const app = new Application({
  mode: 'single',
  env: 'prod',
});

app.listen(9000, '0.0.0.0', () => {
  console.log('Server start on http://0.0.0.0:9000');
});

nextjs
#!/var/lang/node12/bin/node

/*
# Since the HTTP direct function runs based on a Docker image, it must listen to the address 0.0.0.0 and the port must be set to 9000.
*/
const { nextStart } = require('next/dist/cli/next-start');
nextStart(['--port', '9000', '--hostname', '0.0.0.0']);

nuxtjs
#!/var/lang/node12/bin/node

/*
# Since the HTTP direct function runs based on a Docker image, it must listen to the address 0.0.0.0 and the port must be set to 9000.
*/
require('@nuxt/cli')
  .run(['start', '--port', '9000', '--hostname', '0.0.0.0'])
  .catch((error) => {
    require('consola').fatal(error);
    require('exit')(2);
  });

nestjs
#!/bin/bash

# SERVERLESS=1 /var/lang/node12/bin/npm run start -- -e /var/lang/node12/bin/node
SERVERLESS=1 /var/lang/node12/bin/node ./dist/main.js
flask
#!/bin/bash

# Since the HTTP direct function runs based on a Docker image, it must listen to the address 0.0.0.0 and the port must be set to 9000.
/var/lang/python3/bin/python3 app.py
django
#!/bin/bash

# Since the HTTP direct function runs based on a Docker image, it must listen to the address 0.0.0.0 and the port must be set to 9000.
/var/lang/python3/bin/python3 manage.py runserver 0.0.0.0:9000

laravel
#!/bin/bash

#######################################
# Inject Environment Variables in the Serverless Environment
#######################################
# Injecting the SERVERLESS Identifier
export SERVERLESS=1
# Modify the template compilation cache path, as only the /tmp directory is readable and writable in the cloud function.
export VIEW_COMPILED_PATH=/tmp/storage/framework/views
# Modify the session to be stored in memory (array type)
export SESSION_DRIVER=array
# Log Output to stderr
export LOG_CHANNEL=stderr
# Modifying the Application Storage Path
export APP_STORAGE=/tmp/storage

# Initialize Template Cache Directory
mkdir -p /tmp/storage/framework/views

# Since the HTTP direct function runs based on a Docker image, it must listen to the address 0.0.0.0 and the port must be set to 9000.
# Cloud executable file path /var/lang/php7/bin/php
/var/lang/php7/bin/php artisan serve --host 0.0.0.0 --port 9000


中国站 - English