Mysql-Druid连接池配置_druid useunfairlock

1. 背景

DRUID是阿里巴巴开源平台上一个数据库连接池实现，它结合了C3P0、DBCP、PROXOOL等DB池的优点，同时加入了日志监控，可以很好的监控DB池连接和SQL的执行情况，可以说是针对监控而生的DB连接池。

2. 介绍

（1）DataSource中URL重要配置参数（Spring配置前缀：spring.datasource.url）

（2）Druid重要配置参数（Spring配置前缀：spring.datasource.druid）

3. 源码

4. 实战

4.1 数据库连接配置

spring:
  datasource:
    url: jdbc:mysql://127.0.0.1:3306/test?serverTimezone=Asia/Shanghai&useUnicode=true&characterEncoding=UTF8&useSSL=false
    username: root
    password: root
    driver-class-name: com.mysql.cj.jdbc.Driver

4.2 Druid配置

以数据库实例是4核8G为例(待考量)：

spring:
  datasource:
    druid:
      initial-size: 10
      min-idle: 10
      max-active: 25
      max-wait: 60000
      use-unfair-lock: true
      time-between-eviction-runs-millis: 60000
      min-evictable-idle-time-millis: 300000

5. FAQ

5.1 Druid是否有必要在设置了maxWait后，还要强制设置useUnfairLock=true？

设置后，性能会翻倍（见参考资料有赞优化），但是这样做有没有其他问题？

5.2  检测到连接达空闲时间上线时，Druid会物理关闭该连接，那么它不保证最小连接数吗？会重新开启新的连接吗？

(1) com.alibaba.druid.pool.DruidDataSource#getConnectionDirect

public DruidPooledConnection getConnectionDirect(long maxWaitMillis) throws SQLException {
        int notFullTimeoutRetryCnt = 0;
        for (;;) {
            // ...
 
            if (testOnBorrow) {
                // ...
            } else {
                if (poolableConnection.conn.isClosed()) {
                    discardConnection(poolableConnection.holder); // 传入null，避免重复关闭
                    continue;
                }
 
                if (testWhileIdle) {
                    final DruidConnectionHolder holder = poolableConnection.holder;
                    long currentTimeMillis             = System.currentTimeMillis();
                    long lastActiveTimeMillis          = holder.lastActiveTimeMillis;
                    long lastExecTimeMillis            = holder.lastExecTimeMillis;
                    long lastKeepTimeMillis            = holder.lastKeepTimeMillis;
 
                    if (checkExecuteTime
                            && lastExecTimeMillis != lastActiveTimeMillis) {
                        lastActiveTimeMillis = lastExecTimeMillis;
                    }
 
                    if (lastKeepTimeMillis > lastActiveTimeMillis) {
                        lastActiveTimeMillis = lastKeepTimeMillis;
                    }
 
                    long idleMillis                    = currentTimeMillis - lastActiveTimeMillis;
 
                    long timeBetweenEvictionRunsMillis = this.timeBetweenEvictionRunsMillis;
 
                    if (timeBetweenEvictionRunsMillis <= 0) {
                        timeBetweenEvictionRunsMillis = DEFAULT_TIME_BETWEEN_EVICTION_RUNS_MILLIS;
                    }
                    // 如果空闲时间达到阙值
                    if (idleMillis >= timeBetweenEvictionRunsMillis
                            || idleMillis < 0 // unexcepted branch
                            ) {
                        // 检测连接是否有效
                        boolean validate = testConnectionInternal(poolableConnection.holder, poolableConnection.conn);
                        if (!validate) {
                            if (LOG.isDebugEnabled()) {
                                LOG.debug("skip not validate connection.");
                            }
                            // 无效则关闭连接
                            discardConnection(poolableConnection.holder);
                             continue;
                        }
                    }
                }
            }
 
            // ...
    }

(2) 关闭连接后，如果当前存活数小于最小连接数

空信号的处理利用的是AQS的Condition.sign()，这里还不确定会不会开启。Condition忘记了～

public void discardConnection(DruidConnectionHolder holder) {
        if (holder == null) {
            return;
        }
 
        Connection conn = holder.getConnection();
        if (conn != null) {
            JdbcUtils.close(conn);
        }
 
        lock.lock();
        try {
            if (holder.discard) {
                return;
            }
 
            if (holder.active) {
                activeCount--;
                holder.active = false;
            }
            discardCount++;
 
            holder.discard = true;
 
            if (activeCount <= minIdle) {
                emptySignal();
            }
        } finally {
            lock.unlock();
        }
    }

6. 参考资料

【有赞DB连接池性能优化】

【数据库连接池的大小配置】