yipai-tech/AItst

Fork 0

Files

guoyoudu 505cfe929d Initial commit for tst

2026-02-08 14:33:45 +08:00

16 KiB

Raw Blame History

AI监控系统 - BT离线部署指南

📋 概述

本指南介绍如何使用BitTorrent（BT）技术在无网络环境中部署AI监控系统到昇腾服务器，包括种子制作、P2P传输和离线安装。

🔧 BT部署原理

工作流程

有网络环境 → 制作种子 → 分发种子 → P2P下载 → 离线安装

优势

高速传输：多节点并行下载
带宽节省：P2P共享减少服务器压力
断点续传：网络中断后可继续传输
完整性校验：自动验证文件完整性

🌱 制作BT种子

1. 准备完整部署包

#!/bin/bash
# create_deployment_package.sh

echo "=== 制作AI监控系统部署包 ==="

# 创建部署目录结构
DEPLOY_DIR="AIMonitor_deploy_v1.0"
mkdir -p $DEPLOY_DIR/{packages,drivers,models,configs,scripts}

# 1. 复制项目文件
echo "1. 复制项目文件..."
cp -r *.py *.yaml *.md $DEPLOY_DIR/
cp -r requirements.txt $DEPLOY_DIR/

# 2. 下载Python依赖包
echo "2. 下载Python依赖包..."
pip download -r requirements.txt -d $DEPLOY_DIR/packages/ \
    --platform linux_x86_64 --only-binary=:all:

# 3. 下载昇腾驱动和工具包
echo "3. 下载昇腾相关包..."
cd $DEPLOY_DIR/drivers/

# Atlas 300I驱动
wget https://ascend-repo.huawei.com/Atlas%20200I%20DK/Ascend-hdk-23.0.0-ubuntu20.04.aarch64.run

# CANN工具包
wget https://ascend-repo.huawei.com/CANN/CANN%205.0.2/Ascend-cann-toolkit_5.0.2_linux-aarch64.run

# 4. 准备模型文件
echo "4. 准备AI模型..."
cp -r YOLO_Weight/ $DEPLOY_DIR/models/

# 5. 创建配置文件模板
echo "5. 创建配置模板..."
cat > $DEPLOY_DIR/configs/config_template.yaml << EOF
cameras:
  - id: 1
    name: "摄像头1"
    rtsp_url: "rtsp://admin:password@192.168.1.100:554/stream1"
  - id: 2
    name: "摄像头2" 
    rtsp_url: "rtsp://admin:password@192.168.1.101:554/stream1"
EOF

# 6. 创建安装脚本
echo "6. 创建安装脚本..."
cat > $DEPLOY_DIR/scripts/install.sh << 'EOF'
#!/bin/bash
echo "=== AI监控系统安装 ==="

# 安装系统依赖
sudo rpm -ivh packages/*.rpm 2>/dev/null || sudo dpkg -i packages/*.deb

# 安装Python依赖
python3 -m pip install packages/*.whl --no-index --find-links=packages/

# 安装昇腾驱动
sudo bash drivers/Ascend-hdk-*.run --silent

# 安装CANN工具包
sudo bash drivers/Ascend-cann-toolkit-*.run --silent

# 配置环境
echo "source /usr/local/Ascend/ascend-toolkit/set_env.sh" >> ~/.bashrc

echo "安装完成！"
EOF

chmod +x $DEPLOY_DIR/scripts/install.sh

# 7. 创建启动脚本
cat > $DEPLOY_DIR/scripts/start.sh << 'EOF'
#!/bin/bash
cd /opt/AIMonitor
source ~/.bashrc
python3 rtsp_service_ws.py &
python3 static_server.py &
EOF

chmod +x $DEPLOY_DIR/scripts/start.sh

# 8. 创建说明文档
cat > $DEPLOY_DIR/README.txt << EOF
AI监控系统离线部署包 v1.0

安装步骤：
1. 运行 sudo ./scripts/install.sh
2. 配置 configs/config.yaml
3. 运行 ./scripts/start.sh

技术支持：请参考 deploy_升腾.md
EOF

# 9. 压缩部署包
echo "7. 压缩部署包..."
tar -czf AIMonitor_deploy_v1.0.tar.gz $DEPLOY_DIR/

echo "✓ 部署包制作完成: AIMonitor_deploy_v1.0.tar.gz"

2. 制作BT种子

#!/bin/bash
# create_torrent.sh

echo "=== 制作BT种子 ==="

# 安装BT制作工具
pip install transmission-rpc

# 创建种子文件
transmission-create -o AIMonitor_deploy_v1.0.torrent \
    -t "https://tracker1.example.com:6969/announce" \
    -t "https://tracker2.example.com:6969/announce" \
    -c "AI监控系统离线部署包 v1.0 - 昇腾服务器专用" \
    AIMonitor_deploy_v1.0.tar.gz

# 创建带校验和的种子
transmission-show -i AIMonitor_deploy_v1.0.torrent

echo "✓ BT种子创建完成: AIMonitor_deploy_v1.0.torrent"

3. 创建私有Tracker

#!/bin/bash
# setup_tracker.sh

echo "=== 搭建私有BT Tracker ==="

# 使用opentracker作为tracker服务器
docker run -d \
  --name bt-tracker \
  -p 6969:6969 \
  -v $(pwd)/tracker.conf:/etc/opentracker/tracker.conf \
  --restart unless-stopped \
  prologic/opentracker

# 创建tracker配置
cat > tracker.conf << EOF
listen.ipv4_addr = 0.0.0.0
listen.port = 6969
daemon = 0
access.log = access.log
error.log = error.log
debug.log = debug.log
EOF

echo "✓ Private Tracker started on port 6969"

🚀 P2P部署方案

方案一：服务器间P2P传输

1. 在有网络的服务器上

#!/bin/bash
# seed_deployment.sh

echo "=== 启动种子分享 ==="

# 1. 创建完整部署包
./create_deployment_package.sh

# 2. 制作BT种子
./create_torrent.sh

# 3. 启动Transmission作为种子服务器
transmission-daemon
transmission-remote -a AIMonitor_deploy_v1.0.torrent \
    -w ./ \
    --find

echo "✓ 开始分享AI监控系统部署包"
echo "Torrent Info:"
transmission-show -i AIMonitor_deploy_v1.0.torrent

2. 在目标昇腾服务器上

#!/bin/bash
# download_deployment.sh

echo "=== 下载AI监控系统部署包 ==="

# 1. 安装下载工具
pip install transmission-cli

# 2. 从种子文件下载
transmission-cli -f AIMonitor_deploy_v1.0.torrent \
    -w /tmp/ \
    --seedlimit 0

# 3. 校验文件完整性
echo "校验下载文件..."
if [ -f "/tmp/AIMonitor_deploy_v1.0.tar.gz" ]; then
    echo "✓ 文件下载完成"
    
    # 4. 解压并安装
    cd /tmp
    tar -xzf AIMonitor_deploy_v1.0.tar.gz
    cd AIMonitor_deploy_v1.0
    
    # 5. 运行安装脚本
    sudo ./scripts/install.sh
    
    echo "✓ AI监控系统安装完成"
else
    echo "✗ 文件下载失败"
fi

方案二：多节点分布式部署

1. 主节点（有网络）

#!/bin/bash
# master_node.sh

echo "=== 主节点配置 ==="

# 1. 启动Tracker服务
docker-compose up -d

# 2. 创建多个种子版本
python3 create_multi_version_torrents.py

# 3. 启动种子服务
transmission-daemon

# 添加所有种子
for torrent in *.torrent; do
    transmission-remote -a "$torrent" -w ./
done

echo "✓ 主节点启动完成"

2. 从节点（无网络）

#!/bin/bash
# slave_node.sh

echo "=== 从节点部署 ==="

# 1. 连接到主节点Tracker
TRACKER_URL="http://master-server:6969/announce"

# 2. 下载种子文件
curl -O http://master-server:torrents/AIMonitor_deploy_v1.0.torrent

# 3. 多线程下载
transmission-cli \
    -f AIMonitor_deploy_v1.0.torrent \
    -w /tmp/ \
    --peer-info \
    --encryption-preferred

# 4. 验证并安装
if [ -f "/tmp/AIMonitor_deploy_v1.0.tar.gz" ]; then
    cd /tmp && tar -xzf AIMonitor_deploy_v1.0.tar.gz
    cd AIMonitor_deploy_v1.0
    sudo ./scripts/install.sh
else
    echo "下载失败，请检查网络连接"
fi

🛠️ BT部署工具

1. 自动化部署脚本

#!/usr/bin/env python3
# bt_deployment_tool.py

import os
import sys
import hashlib
import subprocess
from pathlib import Path

class BTDeploymentTool:
    def __init__(self):
        self.deployment_dir = "AIMonitor_bt_deploy"
        self.torrent_file = None
        
    def create_deployment_package(self):
        """创建部署包"""
        print("创建部署包...")
        
        # 创建目录结构
        Path(self.deployment_dir).mkdir(exist_ok=True)
        Path(f"{self.deployment_dir}/packages").mkdir(exist_ok=True)
        Path(f"{self.deployment_dir}/scripts").mkdir(exist_ok=True)
        
        # 复制必要文件
        essential_files = [
            "rtsp_service_ws.py", "static_server.py", "monitor_gui.py",
            "npu_yolo_onnx.py", "config.yaml", "requirements.txt"
        ]
        
        for file in essential_files:
            if os.path.exists(file):
                os.system(f"cp {file} {self.deployment_dir}/")
        
        print(f"✓ 部署包创建完成: {self.deployment_dir}")
        
    def create_torrent(self, tracker_urls=None):
        """创建BT种子"""
        print("创建BT种子...")
        
        tracker_urls = tracker_urls or [
            "https://tracker1.example.com:6969/announce",
            "https://tracker2.example.com:6969/announce"
        ]
        
        # 使用libtorrent创建种子
        import libtorrent as lt
        
        # 创建torrent
        fs = lt.file_storage()
        lt.add_files(fs, self.deployment_dir)
        
        t = lt.create_torrent(fs)
        t.add_tracker(tracker_urls[0])
        t.set_creator("AI监控系统部署工具")
        t.set_comment("昇腾AI监控系统离线部署包")
        
        # 生成torrent文件
        self.torrent_file = f"{self.deployment_dir}.torrent"
        with open(self.torrent_file, 'wb') as f:
            f.write(lt.bencode(t.generate()))
        
        print(f"✓ BT种子创建完成: {self.torrent_file}")
        return self.torrent_file
    
    def calculate_checksum(self, file_path):
        """计算文件校验和"""
        sha256_hash = hashlib.sha256()
        with open(file_path, "rb") as f:
            for chunk in iter(lambda: f.read(4096), b""):
                sha256_hash.update(chunk)
        return sha256_hash.hexdigest()
    
    def deploy_from_torrent(self, torrent_path, download_dir="/tmp"):
        """从种子部署"""
        print(f"从种子部署: {torrent_path}")
        
        # 使用transmission下载
        cmd = [
            "transmission-cli", "-f", torrent_path,
            "-w", download_dir, "--seedlimit", "0"
        ]
        
        result = subprocess.run(cmd, capture_output=True, text=True)
        
        if result.returncode == 0:
            # 验证下载文件
            expected_file = os.path.join(download_dir, self.deployment_dir + ".tar.gz")
            if os.path.exists(expected_file):
                print("✓ 下载完成，开始安装...")
                
                # 解压并安装
                os.chdir(download_dir)
                os.system(f"tar -xzf {os.path.basename(expected_file)}")
                
                install_script = os.path.join(self.deployment_dir, "scripts", "install.sh")
                if os.path.exists(install_script):
                    os.system(f"sudo {install_script}")
                    print("✓ 部署完成")
                else:
                    print("✗ 安装脚本不存在")
            else:
                print("✗ 下载文件不完整")
        else:
            print(f"✗ 下载失败: {result.stderr}")

def main():
    tool = BTDeploymentTool()
    
    if len(sys.argv) < 2:
        print("使用方法:")
        print("  python3 bt_deployment_tool.py create   # 创建部署包")
        print("  python3 bt_deployment_tool.py deploy   # 从种子部署")
        return
    
    command = sys.argv[1]
    
    if command == "create":
        tool.create_deployment_package()
        torrent_file = tool.create_torrent()
        print(f"\n种子文件: {torrent_file}")
        
    elif command == "deploy":
        if len(sys.argv) < 3:
            print("请提供种子文件路径")
            return
        tool.deploy_from_torrent(sys.argv[2])
    
    else:
        print(f"未知命令: {command}")

if __name__ == "__main__":
    main()

2. Docker容器化BT部署

# Dockerfile
FROM ubuntu:20.04

RUN apt-get update && apt-get install -y \
    python3-pip \
    transmission-cli \
    curl \
    tar \
    && rm -rf /var/lib/apt/lists/*

RUN pip3 install transmission-rpc libtorrent

WORKDIR /app

COPY . /app/

CMD ["python3", "bt_deployment_tool.py"]

# docker-compose.yml
version: '3.8'

services:
  bt-tracker:
    image: prologic/opentracker
    ports:
      - "6969:6969"
    volumes:
      - ./tracker.conf:/etc/opentracker/tracker.conf
    
  bt-seeder:
    build: .
    volumes:
      - ./deployment:/app/deployment
    command: ["python3", "bt_deployment_tool.py", "create"]
    depends_on:
      - bt-tracker
      
  bt-downloader:
    build: .
    volumes:
      - ./downloads:/app/downloads
    command: ["python3", "bt_deployment_tool.py", "deploy", "/torrents/AIMonitor_deploy_v1.0.torrent"]
    depends_on:
      - bt-tracker

🔐 安全配置

1. 加密传输

# 启用加密传输
transmission-cli \
    --encryption-preferred \
    --encryption-required \
    -f deployment.torrent

2. 访问控制

# tracker访问控制
ALLOWED_PEERS = [
    "192.168.1.0/24",
    "10.0.0.0/8",
    "172.16.0.0/12"
]

def is_peer_allowed(peer_ip):
    """检查对等节点是否被允许"""
    import ipaddress
    peer_addr = ipaddress.ip_address(peer_ip)
    
    for network in ALLOWED_PEERS:
        if peer_addr in ipaddress.ip_network(network):
            return True
    return False

3. 完整性验证

#!/bin/bash
# verify_deployment.sh

TORRENT_FILE=$1
DEPLOYMENT_FILE=$2

echo "验证部署包完整性..."

# 1. 验证torrent信息
transmission-show -i $TORRENT_FILE > torrent_info.txt

# 2. 计算实际文件校验和
ACTUAL_HASH=$(sha256sum $DEPLOYMENT_FILE | cut -d' ' -f1)

# 3. 验证文件大小
EXPECTED_SIZE=$(grep "Size:" torrent_info.txt | awk '{print $2}')
ACTUAL_SIZE=$(stat -c%s $DEPLOYMENT_FILE)

if [ "$EXPECTED_SIZE" = "$ACTUAL_SIZE" ]; then
    echo "✓ 文件大小正确"
else
    echo "✗ 文件大小不匹配"
    exit 1
fi

echo "✓ 部署包验证通过"

📊 监控和管理

1. BT传输监控

#!/usr/bin/env python3
# bt_monitor.py

import time
import transmission_rpc
from datetime import datetime

class BTMonitor:
    def __init__(self, host='localhost', port=9091):
        self.tc = transmission_rpc.Client(host=host, port=port)
    
    def monitor_download(self, torrent_hash):
        """监控下载进度"""
        while True:
            try:
                torrent = self.tc.get_torrent(torrent_hash)
                progress = torrent.progress
                status = torrent.status
                
                print(f"[{datetime.now()}] 进度: {progress:.2f}%, 状态: {status}")
                
                if status == 'seeding':
                    print("下载完成")
                    break
                    
            except Exception as e:
                print(f"监控错误: {e}")
                
            time.sleep(10)
    
    def list_active_torrents(self):
        """列出活跃的种子"""
        torrents = self.tc.get_torrents()
        for torrent in torrents:
            print(f"{torrent.name}: {torrent.progress:.2f}%")

if __name__ == "__main__":
    monitor = BTMonitor()
    monitor.list_active_torrents()

2. 自动重试机制

#!/bin/bash
# auto_retry_download.sh

TORRENT_FILE=$1
MAX_RETRIES=5
RETRY_DELAY=60

for ((i=1; i<=MAX_RETRIES; i++)); do
    echo "尝试下载 (第 $i 次)..."
    
    if transmission-cli -f $TORRENT_FILE -w /tmp/ --seedlimit 0; then
        echo "✓ 下载成功"
        exit 0
    else
        echo "✗ 下载失败，${RETRY_DELAY}秒后重试..."
        sleep $RETRY_DELAY
    fi
done

echo "✗ 达到最大重试次数，下载失败"
exit 1

🎯 部署流程

完整部署步骤

1. 准备阶段（有网络环境）

# 1. 创建部署包
./create_deployment_package.sh

# 2. 制作BT种子
./create_torrent.sh

# 3. 启动种子服务
./seed_deployment.sh

2. 传输阶段（P2P）

# 1. 分发种子文件
scp AIMonitor_deploy_v1.0.torrent user@target-server:/tmp/

# 2. 在目标服务器下载
./download_deployment.sh

3. 安装阶段（离线环境）

# 1. 解压部署包
tar -xzf AIMonitor_deploy_v1.0.tar.gz

# 2. 运行安装脚本
sudo ./scripts/install.sh

# 3. 配置系统
./scripts/configure.sh

# 4. 启动服务
./scripts/start.sh

验证部署结果

# 1. 检查服务状态
systemctl status aimonitor

# 2. 验证端口监听
netstat -tulpn | grep -E "(8765|5000)"

# 3. 测试NPU功能
npu-smi

# 4. 启动GUI界面
python3 monitor_gui.py

文档版本: v1.0
适用平台: 昇腾Atlas系列服务器
网络要求: 支持P2P传输的内网环境

16 KiB Raw Blame History Unescape Escape

AI监控系统 - BT离线部署指南

📋 概述

🔧 BT部署原理

工作流程

优势

🌱 制作BT种子

1. 准备完整部署包

2. 制作BT种子

3. 创建私有Tracker

🚀 P2P部署方案

方案一：服务器间P2P传输

1. 在有网络的服务器上

2. 在目标昇腾服务器上

方案二：多节点分布式部署

1. 主节点（有网络）

2. 从节点（无网络）

🛠️ BT部署工具

1. 自动化部署脚本

2. Docker容器化BT部署

🔐 安全配置

1. 加密传输

2. 访问控制

3. 完整性验证

📊 监控和管理

1. BT传输监控

2. 自动重试机制

🎯 部署流程

完整部署步骤

1. 准备阶段（有网络环境）

2. 传输阶段（P2P）

3. 安装阶段（离线环境）

验证部署结果

16 KiB

Raw Blame History