mirror of
https://github.com/nonebot/nonebot2.git
synced 2024-12-11 14:05:43 +08:00
231 lines
6.5 KiB
Markdown
231 lines
6.5 KiB
Markdown
---
|
||
sidebar_position: 11
|
||
description: 部署你的机器人
|
||
---
|
||
|
||
# 部署
|
||
|
||
在编写完成后,你需要部署你的机器人来使得用户能够使用它。通常,会将机器人部署在服务器上,来保证服务持久运行。
|
||
|
||
在开发时机器人运行的环境称为开发环境,而在部署后机器人运行的环境称为生产环境。与开发环境不同的是,在生产环境中,开发者通常不能随意地修改/添加/删除代码,开启或停止服务。
|
||
|
||
## 部署前准备
|
||
|
||
在生产环境中,为确保机器人能够正常运行,你需要固定你的依赖库版本。下面提供了几种常见的文件格式与生成方式:
|
||
|
||
- `poetry.lock`
|
||
|
||
[poetry](https://python-poetry.org/) 依赖管理工具使用的 lock 文件,通常会在安装依赖时自动生成,或者使用 `poetry lock` 来生成。
|
||
|
||
- `pdm.lock`
|
||
|
||
[pdm](https://pdm.fming.dev/) 依赖管理工具使用的 lock 文件,通常会在安装依赖时自动生成,或者使用 `pdm lock` 来生成。
|
||
|
||
- `Pipfile.lock`
|
||
|
||
[Pipenv](https://pipenv.pypa.io/en/latest/) 依赖管理工具使用的 lock 文件,通常会在安装依赖时自动生成,或者使用 `pipenv lock` 来生成。
|
||
|
||
- `requirements.txt`
|
||
|
||
如果你未使用任何依赖管理工具,你可以使用 `pip freeze` 来生成这个文件。
|
||
|
||
## 使用 Docker 部署(推荐)
|
||
|
||
请自行参考 [Docker 官方文档](https://docs.docker.com/engine/install/) 安装 Docker。
|
||
|
||
在生产环境安装 [docker-compose](https://docs.docker.com/compose/) 工具以便部署机器人。
|
||
|
||
### 编译镜像与部署配置
|
||
|
||
在项目目录下添加以下两个文件(以 poetry 和 FastAPI 驱动器为例):
|
||
|
||
```dockerfile title=Dockerfile
|
||
FROM python:3.9 as requirements-stage
|
||
|
||
WORKDIR /tmp
|
||
|
||
COPY ./pyproject.toml ./poetry.lock* /tmp/
|
||
|
||
RUN curl -sSL https://install.python-poetry.org -o install-poetry.py
|
||
|
||
RUN python install-poetry.py --yes
|
||
|
||
ENV PATH="${PATH}:/root/.local/bin"
|
||
|
||
RUN poetry export -f requirements.txt --output requirements.txt --without-hashes
|
||
|
||
FROM tiangolo/uvicorn-gunicorn-fastapi:python3.9
|
||
|
||
WORKDIR /app
|
||
|
||
COPY --from=requirements-stage /tmp/requirements.txt /app/requirements.txt
|
||
|
||
RUN pip install --no-cache-dir --upgrade -r requirements.txt
|
||
|
||
RUN rm requirements.txt
|
||
|
||
COPY ./ /app/
|
||
```
|
||
|
||
```yaml title=docker-compose.yml
|
||
version: "3"
|
||
services:
|
||
nonebot:
|
||
build: .
|
||
ports:
|
||
- "8080:8080" # 映射端口到宿主机 宿主机端口:容器端口
|
||
env_file:
|
||
- ".env.prod" # fastapi 使用的环境变量文件
|
||
environment:
|
||
- ENVIRONMENT=prod
|
||
- APP_MODULE=bot:app
|
||
- MAX_WORKERS=1
|
||
network_mode: bridge
|
||
```
|
||
|
||
配置完成后即可使用 `docker-compose up -d` 命令来启动机器人并在后台运行。
|
||
|
||
### CI/CD
|
||
|
||
配合 GitHub Actions 可以完成 CI/CD,在 GitHub 上发布 Release 时自动部署至生产环境。
|
||
|
||
在 [Docker Hub](https://hub.docker.com/) 上创建仓库,并将下方 workflow 文件中高亮行中的仓库名称替换为你的仓库名称。
|
||
|
||
前往项目仓库的 `Settings` > `Secrets` > `actions` 栏目 `New Repository Secret` 添加部署所需的密钥:
|
||
|
||
- `DOCKERHUB_USERNAME`: 你的 Docker Hub 用户名
|
||
- `DOCKERHUB_PASSWORD`: 你的 Docker Hub PAT([创建方法](https://docs.docker.com/docker-hub/access-tokens/))
|
||
- `DEPLOY_HOST`: 部署服务器 IP 地址
|
||
- `DEPLOY_USER`: 部署服务器用户名
|
||
- `DEPLOY_KEY`: 部署服务器私钥
|
||
- `DEPLOY_PATH`: 部署服务器上的项目路径
|
||
|
||
将以下文件添加至项目下的 `.github/workflows/` 目录下:
|
||
|
||
```yaml title=.github/workflows/build.yml {30}
|
||
name: Docker Hub Release
|
||
|
||
on:
|
||
push:
|
||
tags:
|
||
- "v*"
|
||
|
||
jobs:
|
||
docker:
|
||
runs-on: ubuntu-latest
|
||
|
||
steps:
|
||
- name: Checkout
|
||
uses: actions/checkout@v2
|
||
|
||
- name: Setup Docker
|
||
uses: docker/setup-buildx-action@v1
|
||
|
||
- name: Login to DockerHub
|
||
uses: docker/login-action@v1
|
||
with:
|
||
username: ${{ secrets.DOCKERHUB_USERNAME }}
|
||
password: ${{ secrets.DOCKERHUB_PASSWORD }}
|
||
|
||
- name: Generate Tags
|
||
uses: docker/metadata-action@v3
|
||
id: metadata
|
||
with:
|
||
images: |
|
||
{organization}/{repository}
|
||
tags: |
|
||
type=semver,pattern={{version}}
|
||
type=sha
|
||
|
||
- name: Build and Publish
|
||
uses: docker/build-push-action@v2
|
||
with:
|
||
context: .
|
||
push: true
|
||
tags: ${{ steps.metadata.outputs.tags }}
|
||
labels: ${{ steps.metadata.outputs.labels }}
|
||
```
|
||
|
||
```yaml title=.github/workflows/deploy.yml
|
||
name: Deploy
|
||
|
||
on:
|
||
workflow_run:
|
||
workflows:
|
||
- Docker Hub Release
|
||
types:
|
||
- completed
|
||
|
||
jobs:
|
||
deploy:
|
||
runs-on: ubuntu-latest
|
||
if: ${{ github.event.workflow_run.conclusion == 'success' }}
|
||
steps:
|
||
- name: start deployment
|
||
uses: bobheadxi/deployments@v0.6
|
||
id: deployment
|
||
with:
|
||
step: start
|
||
token: ${{ secrets.GITHUB_TOKEN }}
|
||
env: official-bot
|
||
|
||
- name: remote ssh command
|
||
uses: appleboy/ssh-action@master
|
||
env:
|
||
DEPLOY_PATH: ${{ secrets.DEPLOY_PATH }}
|
||
with:
|
||
host: ${{ secrets.DEPLOY_HOST }}
|
||
username: ${{ secrets.DEPLOY_USER }}
|
||
key: ${{ secrets.DEPLOY_KEY }}
|
||
envs: DEPLOY_PATH
|
||
script: |
|
||
cd $DEPLOY_PATH
|
||
docker-compose down
|
||
docker-compose pull
|
||
docker-compose up -d
|
||
|
||
- name: update deployment status
|
||
uses: bobheadxi/deployments@v0.6
|
||
if: always()
|
||
with:
|
||
step: finish
|
||
token: ${{ secrets.GITHUB_TOKEN }}
|
||
status: ${{ job.status }}
|
||
deployment_id: ${{ steps.deployment.outputs.deployment_id }}
|
||
```
|
||
|
||
将上一部分的 `docker-compose.yml` 文件以及 `.env.prod` 配置文件添加至 `DEPLOY_PATH` 目录下,并修改 `docker-compose.yml` 文件中的镜像配置,替换为 Docker Hub 的仓库名称。
|
||
|
||
```diff
|
||
- build: .
|
||
+ image: {organization}/{repository}:latest
|
||
```
|
||
|
||
## 使用 Supervisor 部署
|
||
|
||
参考:[Uvicorn - Supervisor](https://www.uvicorn.org/deployment/#supervisor)
|
||
|
||
```ini
|
||
[supervisord]
|
||
|
||
[fcgi-program:nonebot]
|
||
socket=tcp://localhost:8080
|
||
command=python3 -m uvicorn --fd 0 bot:app
|
||
directory=/path/to/bot
|
||
autorestart=true
|
||
startsecs=10
|
||
startretries=3
|
||
numprocs=1
|
||
process_name=%(program_name)s-%(process_num)d
|
||
stdout_logfile=/path/to/log/nonebot.out.log
|
||
stdout_logfile_maxbytes=2MB
|
||
```
|
||
|
||
:::warning 警告
|
||
请配合虚拟环境使用,如 venv 等,请勿直接在 Linux 服务器系统环境中安装。
|
||
:::
|
||
|
||
## 使用 PM2 部署
|
||
|
||
<!-- TODO -->
|