apache.commons.pool2 使用指南

apache.commons.pool2 使用指南

为什么要使用池

创建对象耗时较长，多线程频繁调用等因素限制了我们不能每次使用时都重新创建对象，使用池化思想将对象放进池内，不同线程使用同一个池来获取对象，极大的减少每次业务的调用时间。

一个线程内多次使用这个对象所建立的连接。

使用场景

• 数据库连接池：HikariCP、Druid
• 自定义业务连接池：耗时的客户端连接远程服务端

如何使用

1、创建对象客户端

package com.ncst.client.mcc;import lombok.extern.slf4j.Slf4j;
import java.util.Objects;/*** @Author: Lisy* @Description: MCC 操作工具类*/
@Slf4j
public class MccClient {public MccClient() {}public void putMc(String key, String value) {if (Objects.isNull(key)) {throw new IllegalArgumentException("The key argument cannot be null");}log.info("{}=={}", key, value);}public String getMc(String key) {if (Objects.isNull(key)) {throw new IllegalArgumentException("The " + key + " argument cannot be null");}return "";}}

2、创建池工厂类

package com.ncst.client.mcc;import org.apache.commons.pool2.PooledObject;
import org.apache.commons.pool2.PooledObjectFactory;
import org.apache.commons.pool2.impl.DefaultPooledObject;/*** @Author: Lisy* @Description: 对象池工厂*/
public class MccClientFactory implements PooledObjectFactory<MccClient> {public MccClientFactory() {}@Overridepublic PooledObject<MccClient> makeObject() {MccClient dccClient = new MccClient();return new DefaultPooledObject<>(dccClient);}@Overridepublic void destroyObject(PooledObject<MccClient> p) {MccClient object = p.getObject();object.close();}@Overridepublic boolean validateObject(PooledObject<MccClient> p) {// 验证对象是否可用的逻辑，如果需要return true;}@Overridepublic void activateObject(PooledObject<MccClient> p) {// 激活对象的逻辑，如果需要}@Overridepublic void passivateObject(PooledObject<MccClient> p) {// 钝化对象的逻辑，如果需要}}

3、对象池管理类

用于管理保存的池对象，单例模式

shutdown 方法在不使用时关闭掉对应任务的所有池连接

package com.ncst.client;import com.ncst.client.mcc.MccClient;
import com.ncst.client.mcc.MccClientFactory;
import org.apache.commons.pool2.impl.GenericObjectPool;
import org.apache.commons.pool2.impl.GenericObjectPoolConfig;import java.util.Objects;
import java.util.concurrent.ConcurrentHashMap;/*** @Author: Lisy* @Description: 对象池管理*/
public class MccClientPoolManager {private static volatile MccClientPoolManager INSTANCE;private MccClientPoolManager() {}public static MccClientPoolManager getInstance() {if (Objects.isNull(INSTANCE)) {synchronized (MccClientPoolManager.class) {if (Objects.isNull(INSTANCE)) {INSTANCE = new MccClientPoolManager();}}}return INSTANCE;}/*** key-jobUuid, value-每一个任务池*/private final ConcurrentHashMap<String, GenericObjectPool<MccClient>> POOL_MAP = new ConcurrentHashMap<>();public GenericObjectPool<MccClient> getPool(String jobId, GenericObjectPoolConfig<MccClient> poolConfig) {return POOL_MAP.computeIfAbsent(jobId, k -> new GenericObjectPool<>(new MccClientFactory(), poolConfig));}public void shutdown(String jobId) {GenericObjectPool<MccClient> dccClientGenericObjectPool = POOL_MAP.get(jobId);if (Objects.nonNull(dccClientGenericObjectPool)) {dccClientGenericObjectPool.close();}}
}

4、使用

通过多线程来模拟实际使用场景，如果使用默认poolConfig，那么最大池大小为8个，最多创建8个对象。

通过CountDownLatch 来等待所有任务完成

通过 MccClientPoolManager.getInstance().shutdown(“key”)； 来关闭池

“key” 可根据业务逻辑自行定义

 public static void main(String[] args) throws InterruptedException {int num = 20;CountDownLatch latch = new CountDownLatch(num);for (int i = 0; i < num; i++) {int finalI = i;CompletableFuture.runAsync(() -> {try {GenericObjectPool<MccClient> pool2 = MccClientPoolManager.getInstance().getPool("key", new GenericObjectPoolConfig<>());MccClient mccClient = null;try {mccClient = pool2.borrowObject();mccClient.putMc("" + finalI, "value" + finalI);} finally {if (Objects.nonNull(mccClient)) {pool2.returnObject(mccClient);}}} catch (Exception e) {logger.error(e.getMessage(), e);} finally {latch.countDown();}});}latch.await();MccClientPoolManager.getInstance().shutdown("key");}

源码解析-create()

510：private PooledObject<T> create() throws Exception
创建对象时锁住makeObjectCountLock，判断当前已创建对象是否大于设置的池大小，如果未超过则create==true超过则判断已有的池大小是否为0，如果为0返回create==false，
不为0等待正在创建的对象，执行makeObjectCountLock.waitcreate 为true则创建对象,执行 public PooledObject<T> makeObject()创建完对象后判断是否为空，如果为空将已创建对象 cretaeCount--，并抛出异常验证对象是否可用，如果不可用返回null，并cretaeCount--在finally 中锁住 makeObjectCount lock，将makeObjectCount--，唤醒锁中所有等待的对象将对象放入allObjects map中