python+docker实现分布式存储的demo

test.py代码

#test.py
from flask import Flask, request, jsonify
import requests
import sys
import threadingapp = Flask(__name__)# 存储数据
data_store = {}# 节点列表，通过环境变量传入
nodes = []
current_node = None@app.route('/set', methods=['POST'])
def set_value():key = request.json.get('key')value = request.json.get('value')data_store[key] = value# 同步到其他节点for node in nodes:if node != current_node:try:requests.post(f'http://{node}/sync', json={'key': key, 'value': value})except requests.exceptions.RequestException:pass  # 可以添加更好的错误处理return jsonify({'status': 'success'}), 200@app.route('/get', methods=['GET'])
def get_value():key = request.args.get('key')value = data_store.get(key, None)return jsonify({'value': value}), 200@app.route('/sync', methods=['POST'])
def sync():key = request.json.get('key')value = request.json.get('value')data_store[key] = valuereturn jsonify({'status': 'synced'}), 200@app.route('/nodes', methods=['GET'])
def get_nodes():return jsonify({'nodes': nodes}), 200def run_app(port):app.run(host='0.0.0.0', port=port)if __name__ == '__main__':# 从环境变量获取节点信息current_node = sys.argv[1]  # 例如：'node1:5000'nodes = sys.argv[2].split(',')  # 例如：'node1:5000,node2:5001,node3:5002'port = int(current_node.split(':')[1])run_app(port)

Dockerfile:

# Dockerfile
FROM python:3.9-slimWORKDIR /appCOPY test.py /app/test.py
COPY requirements.txt /app/requirements.txtRUN pip install --no-cache-dir -r requirements.txtEXPOSE 5000CMD ["python", "test.py"]

docker-compose.yml

version: '3'services:node1:build: .container_name: node1ports:- "5000:5000"environment:- NODE_NAME=node1command: ["python", "test.py", "node1:5000", "node1:5000,node2:5000,node3:5000"]networks:- dnetworknode2:build: .container_name: node2ports:- "5001:5000"environment:- NODE_NAME=node2command: ["python", "test.py", "node2:5000", "node1:5000,node2:5000,node3:5000"]networks:- dnetworknode3:build: .container_name: node3ports:- "5002:5000"environment:- NODE_NAME=node3command: ["python", "test.py", "node3:5000", "node1:5000,node2:5000,node3:5000"]networks:- dnetworknetworks:dnetwork:driver: bridge

requirements.txt

Flask
requests

上面的demo是用docker-compose启动三个容器（这三个容器的镜像都是由同一个Dockerfile实现的），然后实现的功能就是，比如我往容器一存东西，然后可以在容器2，容器3查看到

curl测试

curl -X POST -H "Content-Type: application/json" \-d '{"key": "foo", "value": "bar"}' \http://localhost:5000/setcurl http://localhost:5002/get?key=foo

然后我们以后还可以扩展，将docker-compose启动的镜像换为三个不同的镜像来完成你的项目需求（例如分布式机器学习，训练模型等等），整体的demo是搭起来了，后面你的具体需求就需要你自己修改实现了