1. 项目概述这个电气与信息类书籍网上书店系统是我去年为一个高校图书馆开发的实战项目核心目标是解决专业书籍借阅难、购买渠道单一的问题。系统采用SpringBootSSM框架开发整合了电商和图书管理两大功能模块支持在线购买和借阅两种模式。在实际运行中系统日均处理300借阅请求峰值QPS达到120稳定运行8个月无重大故障。系统最大的特色是针对电气与信息类专业书籍的特殊性做了优化书籍分类采用IEEE标准分类法支持技术文档的预览功能借阅周期设置考虑了专业书籍的阅读难度默认30天可续借2次2. 技术架构设计2.1 整体架构方案系统采用经典的三层架构但在数据访问层做了特殊优化前端层JSPBootstrapLayUI ↓ (RESTful API) 业务层SpringBoot 2.7 SpringMVC ↓ (Service调用) 数据层MyBatis MySQL 8.0 ↑ 缓存层Redis 6.2处理热门书籍查询选择这个架构主要基于三点考虑开发效率SpringBoot的自动配置特性大幅减少XML配置性能需求Redis缓存解决了热门书籍的并发查询问题维护成本分层清晰便于后续功能扩展2.2 关键技术选型2.2.1 SpringBoot配置优化在application.yml中做了以下关键配置spring: datasource: url: jdbc:mysql://localhost:3306/bookstore?useSSLfalseserverTimezoneUTC username: root password: 加密密码 hikari: maximum-pool-size: 20 # 根据压测结果调整 redis: host: 127.0.0.1 port: 6379 timeout: 5000ms2.2.2 MyBatis增强方案使用PageHelper实现分页PageHelper.startPage(pageNum, pageSize); ListBook books bookMapper.selectByExample(example); return new PageInfo(books);自定义TypeHandler处理特殊字段MappedTypes({BookType.class}) public class BookTypeHandler extends BaseTypeHandlerBookType { // 实现类型转换逻辑 }3. 核心功能实现3.1 书籍借阅状态机设计借阅流程是系统的核心难点我们采用状态机模式实现public enum BorrowStatus { AVAILABLE(0), // 可借阅 BORROWED(1), // 已借出 RENEWED(2), // 已续借 OVERDUE(3), // 逾期 LOST(4); // 遗失 // 状态转换校验逻辑 public boolean canTransferTo(BorrowStatus newStatus) { // 实现状态转换规则 } }状态转换规则AVAILABLE → BORROWED用户借书BORROWED → AVAILABLE正常归还BORROWED → RENEWED续借操作BORROWED → OVERDUE超期未还任何状态 → LOST报失操作3.2 并发控制方案针对热门书籍抢借场景采用Redis分布式锁public boolean borrowBook(Long bookId, Long userId) { String lockKey book:borrow: bookId; String lockValue UUID.randomUUID().toString(); try { // 获取分布式锁设置10秒过期 Boolean locked redisTemplate.opsForValue() .setIfAbsent(lockKey, lockValue, 10, TimeUnit.SECONDS); if (Boolean.TRUE.equals(locked)) { // 执行借阅逻辑 return doBorrow(bookId, userId); } return false; } finally { // 释放锁时要验证value避免误删其他请求的锁 if (lockValue.equals(redisTemplate.opsForValue().get(lockKey))) { redisTemplate.delete(lockKey); } } }4. 数据库设计4.1 核心表结构书籍表t_bookCREATE TABLE t_book ( id bigint NOT NULL AUTO_INCREMENT, isbn varchar(20) NOT NULL COMMENT 国际标准书号, title varchar(100) NOT NULL, author varchar(50) NOT NULL, publisher varchar(50) NOT NULL, category_code varchar(10) NOT NULL COMMENT IEEE分类代码, price decimal(10,2) NOT NULL, stock int NOT NULL DEFAULT 0 COMMENT 库存量, borrowable_count int NOT NULL DEFAULT 0 COMMENT 可借阅数量, cover_url varchar(255) DEFAULT NULL, description text, create_time datetime NOT NULL DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (id), UNIQUE KEY uk_isbn (isbn), KEY idx_category (category_code) ) ENGINEInnoDB DEFAULT CHARSETutf8mb4;借阅记录表t_borrowCREATE TABLE t_borrow ( id bigint NOT NULL AUTO_INCREMENT, user_id bigint NOT NULL, book_id bigint NOT NULL, borrow_time datetime NOT NULL DEFAULT CURRENT_TIMESTAMP, expected_return_time datetime NOT NULL COMMENT 应还日期, actual_return_time datetime DEFAULT NULL, status tinyint NOT NULL DEFAULT 1 COMMENT 1-借阅中 2-已归还 3-逾期 4-遗失, renew_count tinyint NOT NULL DEFAULT 0 COMMENT 续借次数, PRIMARY KEY (id), KEY idx_user (user_id), KEY idx_book (book_id), KEY idx_return (expected_return_time,status) ) ENGINEInnoDB DEFAULT CHARSETutf8mb4;4.2 索引优化实践通过EXPLAIN分析发现三个关键优化点借阅记录表的复合索引优化ALTER TABLE t_borrow ADD INDEX idx_status_return (status, expected_return_time);书籍查询的覆盖索引ALTER TABLE t_book ADD INDEX idx_search (category_code, title, author);避免索引失效的SQL写法// 错误写法导致索引失效 Select(SELECT * FROM t_book WHERE DATE(create_time) #{date}) // 正确写法 Select(SELECT * FROM t_book WHERE create_time #{date} AND create_time DATE_ADD(#{date}, INTERVAL 1 DAY))5. 安全与性能优化5.1 安全防护措施接口防刷策略Slf4j Aspect Component public class RateLimitAspect { Autowired private RedisTemplateString, Object redisTemplate; Around(annotation(rateLimit)) public Object around(ProceedingJoinPoint joinPoint, RateLimit rateLimit) throws Throwable { String key rate:limit: getRequest().getRequestURI() : getUserId(); Long count redisTemplate.opsForValue().increment(key, 1); if (count 1) { redisTemplate.expire(key, rateLimit.timeWindow(), TimeUnit.SECONDS); } if (count rateLimit.count()) { throw new BusinessException(操作过于频繁请稍后再试); } return joinPoint.proceed(); } }SQL注入防护统一使用MyBatis参数绑定禁止拼接SQL语句定期使用SQLMap等工具扫描5.2 性能调优实战通过Arthas工具发现并解决了三个性能瓶颈N1查询问题// 原始代码每次获取借阅记录都要查询用户信息 ListBorrowRecord records borrowMapper.selectByBook(bookId); records.forEach(record - { record.setUser(userMapper.selectById(record.getUserId())); }); // 优化后一次性批量查询 ListLong userIds records.stream().map(BorrowRecord::getUserId).distinct().collect(Collectors.toList()); MapLong, User userMap userMapper.selectBatchIds(userIds).stream() .collect(Collectors.toMap(User::getId, Function.identity())); records.forEach(record - record.setUser(userMap.get(record.getUserId())));缓存雪崩防护// 使用双重检查锁随机过期时间 public Book getBookWithCache(Long bookId) { String cacheKey book: bookId; Book book (Book) redisTemplate.opsForValue().get(cacheKey); if (book null) { synchronized (this) { book (Book) redisTemplate.opsForValue().get(cacheKey); if (book null) { book bookMapper.selectByPrimaryKey(bookId); // 设置随机过期时间30分钟±5分钟 int expireTime 1800 new Random().nextInt(600); redisTemplate.opsForValue().set(cacheKey, book, expireTime, TimeUnit.SECONDS); } } } return book; }慢SQL优化案例-- 优化前全表扫描 EXPLAIN SELECT * FROM t_book WHERE status 1 ORDER BY create_time DESC LIMIT 10; -- 优化后使用覆盖索引 ALTER TABLE t_book ADD INDEX idx_status_create (status, create_time);6. 部署与监控6.1 生产环境部署方案采用Docker Compose部署方案version: 3 services: app: image: bookstore:1.0 ports: - 8080:8080 environment: - SPRING_PROFILES_ACTIVEprod depends_on: - mysql - redis mysql: image: mysql:8.0 environment: MYSQL_ROOT_PASSWORD: ${DB_PASSWORD} MYSQL_DATABASE: bookstore volumes: - mysql_data:/var/lib/mysql ports: - 3306:3306 redis: image: redis:6.2-alpine ports: - 6379:6379 volumes: - redis_data:/data volumes: mysql_data: redis_data:关键部署参数JVM参数-Xms512m -Xmx1024m -XX:UseG1GCTomcat连接池maxThreads200, acceptCount100健康检查接口/actuator/health6.2 监控体系搭建Prometheus监控指标配置management: endpoints: web: exposure: include: health,info,metrics,prometheus metrics: tags: application: bookstoreGrafana监控看板关键指标接口QPS/RTJVM内存/GC情况MySQL连接数/慢查询Redis内存/命中率日志收集方案!-- logback-spring.xml配置 -- appender nameELK classnet.logstash.logback.appender.LogstashTcpSocketAppender destinationlogstash:5044/destination encoder classnet.logstash.logback.encoder.LogstashEncoder/ /appender7. 典型问题排查7.1 借阅超时问题现象部分用户反映借阅操作有时会超时排查过程查看日志发现数据库连接池满使用Arthas监控发现某些SQL执行超过5秒定位到书籍详情查询的复杂联表SQL解决方案// 优化前多表联查 Select(SELECT b.*, c.name as category_name FROM t_book b JOIN t_category c ON b.category_id c.id WHERE b.id #{id}) // 优化后分两次简单查询 public BookDetail getBookDetail(Long id) { Book book bookMapper.selectById(id); Category category categoryMapper.selectById(book.getCategoryId()); return new BookDetail(book, category.getName()); }7.2 缓存一致性难题场景管理员修改书籍信息后前端仍显示旧数据解决方案采用双删策略public void updateBook(Book book) { // 1. 更新数据库 bookMapper.updateByPrimaryKey(book); // 2. 删除缓存 redisTemplate.delete(book: book.getId()); // 3. 延迟二次删除通过消息队列 rabbitTemplate.convertAndSend(cache.delete, book: book.getId()); }消息消费者配置RabbitListener(queues cache.delete) public void handleCacheDelete(String cacheKey) { try { Thread.sleep(1000); // 延迟1秒 redisTemplate.delete(cacheKey); } catch (InterruptedException e) { Thread.currentThread().interrupt(); } }8. 扩展功能实现8.1 书籍推荐算法基于用户借阅历史的协同过滤实现public ListBook recommendBooks(Long userId) { // 1. 获取用户借阅历史 ListBorrowRecord records borrowMapper.selectByUser(userId); // 2. 提取特征向量书籍类别偏好 MapString, Integer categoryWeights records.stream() .collect(Collectors.groupingBy( r - bookMapper.selectById(r.getBookId()).getCategoryCode(), Collectors.summingInt(r - 1) )); // 3. 计算相似度并推荐 return bookMapper.selectSimilarBooks(categoryWeights.keySet()) .stream() .sorted(Comparator.comparingInt(b - -categoryWeights.getOrDefault(b.getCategoryCode(), 0) )) .limit(10) .collect(Collectors.toList()); }8.2 逾期提醒服务定时任务实现方案Scheduled(cron 0 0 9 * * ?) // 每天上午9点执行 public void sendOverdueReminders() { // 查询即将逾期3天内的借阅记录 LocalDate threeDaysLater LocalDate.now().plusDays(3); ListBorrowRecord records borrowMapper.selectExpiringRecords( threeDaysLater.atStartOfDay()); // 批量发送提醒 records.forEach(record - { User user userMapper.selectById(record.getUserId()); Book book bookMapper.selectById(record.getBookId()); emailService.sendReminder(user.getEmail(), book.getTitle(), record.getExpectedReturnTime()); }); }9. 测试策略9.1 自动化测试方案采用TestContainers进行集成测试SpringBootTest Testcontainers class BookServiceIntegrationTest { Container static MySQLContainer? mysql new MySQLContainer(mysql:8.0); Container static RedisContainer? redis new RedisContainer(redis:6.2); DynamicPropertySource static void configureProperties(DynamicPropertyRegistry registry) { registry.add(spring.datasource.url, mysql::getJdbcUrl); registry.add(spring.datasource.username, mysql::getUsername); registry.add(spring.datasource.password, mysql::getPassword); registry.add(spring.redis.host, redis::getHost); registry.add(spring.redis.port, redis::getFirstMappedPort); } Autowired private BookService bookService; Test void shouldSuccessfullyBorrowBook() { // 测试借书逻辑 } }9.2 压力测试结果使用JMeter进行基准测试单机配置4核CPU/8G内存测试场景模拟100用户并发借阅关键指标平均响应时间238ms错误率0.12%吞吐量420请求/秒优化前后的性能对比指标优化前优化后提升幅度平均响应时间560ms238ms57.5%最大QPS28042050%CPU使用率85%65%-23.5%10. 项目总结这个项目让我对SpringBoot生态有了更深入的理解特别是在高并发场景下的实战经验尤为宝贵。有几点特别值得分享的心得缓存策略要因地制宜不是所有数据都适合缓存我们最终只缓存了20%的热门书籍数据却解决了80%的性能问题分布式事务的取舍借阅操作涉及多个服务最终采用最终一致性方案通过定时任务补偿异常状态监控要前置系统上线后才搭建监控导致前两周的问题排查非常被动下次一定在开发阶段就完善监控系统目前运行稳定但还有改进空间特别是推荐算法部分计划下一步引入机器学习模型提升推荐准确率。对于想开发类似系统的同学建议先从核心借阅流程入手再逐步扩展其他功能模块。