#!/usr/bin/env python3 # -*- coding: UTF-8 -*- # 配置项 DEFAULT_DEVICE = "/dev/sda" # 默认检查的设备路径 DEFAULT_CONTROLLER_COUNT = 8 # 默认控制器数量 DEFAULT_LOG_DIR = "/root" # 默认日志目录 MCP_SERVER_URL = "http://10.10.13.143:4527/mcp/v1/submit" # MCP服务器提交接口 CRITICAL_ATTRIBUTES = ["Reallocated_Sector_Ct", "Spin_Retry_Count", "End-to-End_Error", "CRC_Error_Count", "Multi_Zone_Error_Rate"] # 关注的重要SMART属性 import subprocess import sys import os import datetime import re import json import requests class DiskInspection: def __init__(self, device=DEFAULT_DEVICE, controller_count=DEFAULT_CONTROLLER_COUNT, log_dir=DEFAULT_LOG_DIR): self.device = device self.controller_count = controller_count self.log_dir = log_dir self.current_date = datetime.datetime.now().strftime("%y%m%d") self.log_file = os.path.join(log_dir, f"smartctl.{self.current_date}.log") self.md_report = os.path.join(log_dir, f"disk_inspection_report.{self.current_date}.md") self.results = {} def run_smartctl(self, controller_id): """执行smartctl命令检查指定控制器的硬盘信息""" try: command = f"smartctl --all -d megaraid,{controller_id} {self.device}" result = subprocess.run( command, shell=True, check=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE, universal_newlines=True ) # 保存到日志文件 with open(self.log_file, "a") as f: f.write(f"\n\n===== Controller {controller_id} =====\n") f.write(result.stdout) return result.stdout except subprocess.CalledProcessError as e: error_msg = f"执行smartctl命令失败 (Controller {controller_id}): {e.stderr}\n" print(error_msg) # 将错误信息也写入日志 with open(self.log_file, "a") as f: f.write(f"\n\n===== Controller {controller_id} (Error) =====\n") f.write(error_msg) return None def parse_smart_info(self, output, controller_id): """解析smartctl输出的信息，根据实际smartctl_log格式调整""" if not output: return {"error": "No data available"} parsed_info = {} # 提取基本信息 vendor_match = re.search(r"Vendor:\s+(.*)", output) if vendor_match: parsed_info["vendor"] = vendor_match.group(1) product_match = re.search(r"Product:\s+(.*)", output) if product_match: parsed_info["device_model"] = product_match.group(1) revision_match = re.search(r"Revision:\s+(.*)", output) if revision_match: parsed_info["firmware_version"] = revision_match.group(1) serial_match = re.search(r"Serial number:\s+(.*)", output) if serial_match: parsed_info["serial_number"] = serial_match.group(1) # 提取容量信息 capacity_match = re.search(r"User Capacity:\s+(.*?)\s+bytes", output) if capacity_match: parsed_info["capacity"] = capacity_match.group(1) # 提取旋转速率 rotation_match = re.search(r"Rotation Rate:\s+(.*)", output) if rotation_match: parsed_info["rotation_rate"] = rotation_match.group(1) # 提取SMART健康状态 health_match = re.search(r"SMART Health Status:\s+(.*)", output) if health_match: parsed_info["health_status"] = health_match.group(1) # 提取温度信息 temp_match = re.search(r"Current Drive Temperature:\s+([0-9]+)", output) if temp_match: parsed_info["temperature"] = temp_match.group(1) # 提取通电时间 power_on_match = re.search(r"Accumulated power on time, hours:minutes\s+([0-9:]+)", output) if power_on_match: parsed_info["power_on_time"] = power_on_match.group(1) # 提取制造日期 manufactured_match = re.search(r"Manufactured in week (\d+) of year (\d+)", output) if manufactured_match: parsed_info["manufactured_date"] = f"{manufactured_match.group(2)}-W{manufactured_match.group(1)}" # 提取错误计数信息 error_read_match = re.search(r"read:\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+\.\d+)\s+(\d+)", output) if error_read_match: parsed_info["read_errors"] = { "corrected": error_read_match.group(4), "uncorrected": error_read_match.group(7), "data_processed": f"{error_read_match.group(6)} GB" } error_write_match = re.search(r"write:\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+\.\d+)\s+(\d+)", output) if error_write_match: parsed_info["write_errors"] = { "corrected": error_write_match.group(4), "uncorrected": error_write_match.group(7), "data_processed": f"{error_write_match.group(6)} GB" } # 提取SMART自检日志信息 self_test_match = re.search(r"SMART Self-test log\nNum\s+Test\s+Status\s+segment\s+LifeTime\s+LBA_first_err \[SK ASC ASQ\]\n(.*?)(\n\s*\n|$)", output, re.DOTALL) if self_test_match: test_results = [] lines = self_test_match.group(1).strip().split("\n") for line in lines: if line.strip() and not line.startswith(" "): parts = re.split(r"\s+", line.strip()) if len(parts) >= 5: test_info = { "id": parts[0], "type": parts[1], "status": parts[2], "lifetime_hours": parts[4] } test_results.append(test_info) if test_results: parsed_info["self_test_results"] = test_results return parsed_info def generate_md_report(self): """生成Markdown格式的报告，根据实际smartctl_log格式调整""" with open(self.md_report, "w") as f: f.write(f"# 硬盘巡检报告\n\n") f.write(f"**检查日期**: {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n\n") f.write(f"**设备**: {self.device}\n\n") f.write("## 控制器检查结果\n\n") # 添加健康状态概览 healthy_count = sum(1 for cid, info in self.results.items() if "health_status" in info and info["health_status"] == "OK") detected_count = sum(1 for cid, info in self.results.items() if "health_status" in info) f.write(f"### 健康状态概览\n") f.write(f"- 总控制器数: {self.controller_count}\n") f.write(f"- 检测到的控制器数: {detected_count}\n") f.write(f"- 健康控制器数: {healthy_count}\n") f.write(f"- 异常控制器数: {detected_count - healthy_count}\n\n") # 详细信息表格 - 简化版 f.write("### 基本信息概览\n") f.write("| 控制器ID | 厂商 | 设备型号 | 序列号 | 健康状态 | 温度(°C) | 通电时间 |\n") f.write("|----------|------|----------|--------|----------|----------|----------|\n") for controller_id in range(self.controller_count): info = self.results.get(controller_id, {}) if "health_status" in info: # 只显示有数据的控制器 f.write(f"| {controller_id} ") f.write(f"| {info.get('vendor', 'N/A')} ") f.write(f"| {info.get('device_model', 'N/A')} ") f.write(f"| {info.get('serial_number', 'N/A')} ") f.write(f"| {info.get('health_status', 'N/A')} ") f.write(f"| {info.get('temperature', 'N/A')} ") f.write(f"| {info.get('power_on_time', 'N/A')} |\n") # 为每个控制器添加详细信息部分 f.write("\n## 控制器详细信息\n\n") has_issues = False for controller_id in range(self.controller_count): info = self.results.get(controller_id, {}) if "health_status" in info: # 只显示有数据的控制器 f.write(f"### 控制器 {controller_id}\n") f.write(f"- **厂商**: {info.get('vendor', 'N/A')}\n") f.write(f"- **设备型号**: {info.get('device_model', 'N/A')}\n") f.write(f"- **序列号**: {info.get('serial_number', 'N/A')}\n") f.write(f"- **固件版本**: {info.get('firmware_version', 'N/A')}\n") f.write(f"- **容量**: {info.get('capacity', 'N/A')}\n") f.write(f"- **旋转速率**: {info.get('rotation_rate', 'N/A')}\n") f.write(f"- **制造日期**: {info.get('manufactured_date', 'N/A')}\n") f.write(f"- **健康状态**: **{info.get('health_status', 'N/A')}**\n") f.write(f"- **当前温度**: {info.get('temperature', 'N/A')}°C\n") f.write(f"- **累计通电时间**: {info.get('power_on_time', 'N/A')}\n") # 错误计数信息 if "read_errors" in info: read_err = info["read_errors"] f.write("- **读取错误**:\n") f.write(f" - 已纠正: {read_err['corrected']}\n") f.write(f" - 未纠正: {read_err['uncorrected']}\n") f.write(f" - 处理数据量: {read_err['data_processed']}\n") if "write_errors" in info: write_err = info["write_errors"] f.write("- **写入错误**:\n") f.write(f" - 已纠正: {write_err['corrected']}\n") f.write(f" - 未纠正: {write_err['uncorrected']}\n") f.write(f" - 处理数据量: {write_err['data_processed']}\n") # SMART自检日志 if "self_test_results" in info: f.write("- **SMART自检结果**:\n") for test in info["self_test_results"]: f.write(f" - 测试 #{test['id']}: {test['type']} - {test['status']} (运行时间: {test['lifetime_hours']}小时)\n") # 检查是否有异常 if info.get("health_status", "") != "OK": has_issues = True f.write("\n**⚠️ 警告：此控制器状态异常，请及时关注！**\n") f.write("\n") f.write("\n## 异常汇总\n\n") if has_issues: f.write("**发现异常控制器，请关注以下问题：**\n\n") for controller_id in range(self.controller_count): info = self.results.get(controller_id, {}) if info.get("health_status", "") != "OK": f.write(f"- **控制器 {controller_id}**: 状态为 {info.get('health_status', '未知')}\n") f.write("\n") else: f.write("**未发现异常情况。**\n\n") f.write("\n## 建议\n\n") if has_issues: f.write("1. **立即关注异常控制器**的状态变化，并考虑进行进一步测试\n") f.write("2. **备份重要数据**，防止因硬盘故障导致数据丢失\n") f.write("3. **联系技术支持**，评估是否需要更换有问题的硬盘\n") f.write("4. 增加监控频率，密切关注错误计数的变化趋势\n") else: f.write("1. **保持定期检查**硬盘健康状态，建议每两周至少检查一次\n") f.write("2. **继续做好数据备份工作**，采用异地备份和多版本策略\n") f.write("3. **注意监控系统温度变化**，保持良好的散热环境\n") f.write("4. 记录并跟踪硬盘的通电时间和错误计数趋势\n") print(f"Markdown报告已生成: {self.md_report}") return self.md_report def submit_to_mcp(self, report_path): """将报告提交给AI分析（通过MCP服务器）""" try: # 读取报告内容 with open(report_path, 'r') as f: report_content = f.read() # 构建提交给MCP的数据 - 按照服务器要求的格式，timestamp应为Unix时间戳整数 mcp_data = { "data": { "type": "disk_inspection_report", "content": report_content }, "type": "disk_inspection", "timestamp": int(datetime.datetime.now().timestamp()) } print("正在准备通过MCP提交报告给AI分析...") # 创建一个临时JSON文件，用于存储要提交的数据 temp_json = os.path.join(self.log_dir, f"mcp_data_{self.current_date}.json") with open(temp_json, 'w') as f: json.dump(mcp_data, f, ensure_ascii=False, indent=2) print(f"MCP数据已准备好: {temp_json}") # 尝试使用requests库直接向MCP服务器发送请求 try: # MCP服务器配置 print(f"正在向MCP服务器发送请求: {MCP_SERVER_URL}") response = requests.post( MCP_SERVER_URL, json=mcp_data, headers={"Content-Type": "application/json"} ) if response.status_code == 200: result = response.json() print("MCP提交成功！") print(f"AI分析结果: {result.get('result', {}).get('analysis', '无分析结果')}") # 保存AI分析结果到文件 analysis_file = os.path.join(self.log_dir, f"ai_analysis_{self.current_date}.json") with open(analysis_file, 'w') as f: json.dump(result, f, ensure_ascii=False, indent=2) print(f"AI分析结果已保存: {analysis_file}") else: print(f"MCP服务器请求失败: HTTP {response.status_code}") print(f"响应内容: {response.text}") print("请检查MCP服务器是否正常运行") except ImportError: print("未安装requests库，无法直接发送HTTP请求") print("请运行: pip install requests 来安装必要的依赖") print("或者使用以下命令手动提交数据:") print(f"curl -X POST -H 'Content-Type: application/json' -d @{temp_json} {MCP_SERVER_URL}") except Exception as e: print(f"发送请求到MCP服务器时出错: {str(e)}") print("请检查MCP服务器是否正常运行") print("或者使用以下命令手动提交数据:") print(f"curl -X POST -H 'Content-Type: application/json' -d @{temp_json} {MCP_SERVER_URL}") return True except Exception as e: print(f"提交MCP失败: {str(e)}") return False def run_inspection(self): """运行完整的硬盘巡检流程""" print(f"开始硬盘巡检，设备: {self.device}") print(f"日志文件: {self.log_file}") # 清空日志文件（如果存在） if os.path.exists(self.log_file): open(self.log_file, 'w').close() # 检查每个控制器 for controller_id in range(self.controller_count): print(f"检查控制器 {controller_id}...") output = self.run_smartctl(controller_id) if output: self.results[controller_id] = self.parse_smart_info(output, controller_id) # 生成MD报告 report_path = self.generate_md_report() # 提交给MCP self.submit_to_mcp(report_path) print("硬盘巡检完成！") if __name__ == "__main__": # 检查是否以root权限运行 if os.geteuid() != 0: print("错误：请以root权限运行此脚本") sys.exit(1) # 检查是否安装了smartctl try: subprocess.run("which smartctl", shell=True, check=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE) except subprocess.CalledProcessError: print("错误：未找到smartctl工具，请先安装") print("Ubuntu/Debian: apt install smartmontools") print("CentOS/RHEL: yum install smartmontools") sys.exit(1) # 创建并运行巡检实例 inspector = DiskInspection() inspector.run_inspection()