Mysql C3PO:客户端在等待获取资源时超时

当前设置：

Generic configuration, inherited by primary and worker DB
    <property name="maxIdleTime" value="44000"/>
    <property name="idleConnectionTestPeriod" value="30"/>
    <property name="maxStatements" value="50"/>
    <property name="minPoolSize" value="3"/>
    <property name="maxPoolSize" value="3"/>
    <property name="acquireIncrement" value="2"/>
    <property name="checkoutTimeout" value="15000"/>
    <property name="acquireRetryDelay" value="1000"/>

Primary DB specific config
    <property name="acquireRetryAttempts" value="0"/>
    <property name="minPoolSize" value="3"/>
    <property name="maxPoolSize" value="5"/>
    <property name="testConnectionOnCheckout" value="true"/>
    <property name="connectionTesterClassName" value="${Custom.connectionTesterClassName}"/>

服务主机（Java）连接到（JDBC）主数据库（MySQL RW），并有一个工作线程（只读）以确保可靠性。连接管理由C3PO处理

场景：

Generic configuration, inherited by primary and worker DB
    <property name="maxIdleTime" value="44000"/>
    <property name="idleConnectionTestPeriod" value="30"/>
    <property name="maxStatements" value="50"/>
    <property name="minPoolSize" value="3"/>
    <property name="maxPoolSize" value="3"/>
    <property name="acquireIncrement" value="2"/>
    <property name="checkoutTimeout" value="15000"/>
    <property name="acquireRetryDelay" value="1000"/>

Primary DB specific config
    <property name="acquireRetryAttempts" value="0"/>
    <property name="minPoolSize" value="3"/>
    <property name="maxPoolSize" value="5"/>
    <property name="testConnectionOnCheckout" value="true"/>
    <property name="connectionTesterClassName" value="${Custom.connectionTesterClassName}"/>

在翻转的情况下，我需要将其中一个worker提升到primarydb。当前主服务器变为只读，新主服务器升级为读写。我希望可靠地将C3P0连接池自动刷新到新的主数据库。哪一种现在起作用，但在某些情况下会被破坏

技术：服务有一个数据库主机作为cname，它将在flip上更新。在DNS更改传播和新db为RW之前，所有连接都将失败。尝试在连接池上以内联方式设置C3PO重试，直到连接通过RW（收单机构尝试）

现有C3PO配置：

Generic configuration, inherited by primary and worker DB
    <property name="maxIdleTime" value="44000"/>
    <property name="idleConnectionTestPeriod" value="30"/>
    <property name="maxStatements" value="50"/>
    <property name="minPoolSize" value="3"/>
    <property name="maxPoolSize" value="3"/>
    <property name="acquireIncrement" value="2"/>
    <property name="checkoutTimeout" value="15000"/>
    <property name="acquireRetryDelay" value="1000"/>

Primary DB specific config
    <property name="acquireRetryAttempts" value="0"/>
    <property name="minPoolSize" value="3"/>
    <property name="maxPoolSize" value="5"/>
    <property name="testConnectionOnCheckout" value="true"/>
    <property name="connectionTesterClassName" value="${Custom.connectionTesterClassName}"/>

最终日志：

org.hibernate.engine.jdbc.internal.LogicalConnectionImpl: Obtaining JDBC connection
com.mchange.v2.resourcepool.BasicResourcePool: acquire test -- pool is already maxed out. [managed: 5; max: 5]
com.mchange.v2.resourcepool.BasicResourcePool: awaitAvailable(): com.mchange.v2.c3p0.impl.NewPooledConnection@4b52cf0c

com.mchange.v2.sql.SqlUtils: Converting Throwable to SQLException...
com.mchange.v2.resourcepool.TimeoutException: A client timed out while waiting to acquire a resource from com.mchange.v2.resourcepool.BasicResourcePool@43aeb5e0 -- timeout at awaitAvailable()
    at com.mchange.v2.resourcepool.BasicResourcePool.awaitAvailable(BasicResourcePool.java:1467) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.resourcepool.BasicResourcePool.prelimCheckoutResource(BasicResourcePool.java:644) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.resourcepool.BasicResourcePool.checkoutResource(BasicResourcePool.java:554) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.c3p0.impl.C3P0PooledConnectionPool.checkoutAndMarkConnectionInUse(C3P0PooledConnectionPool.java:758) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.c3p0.impl.C3P0PooledConnectionPool.checkoutPooledConnection(C3P0PooledConnectionPool.java:685) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.c3p0.impl.AbstractPoolBackedDataSource.getConnection(AbstractPoolBackedDataSource.java:140) ~[c3p0-0.9.5.1.jar:0.9.5.1]

org.hibernate.engine.jdbc.spi.SqlExceptionHelper: Could not open connection [n/a]
java.sql.SQLException: An attempt by a client to checkout a Connection has timed out.
    at com.mchange.v2.sql.SqlUtils.toSQLException(SqlUtils.java:118) ~[mchange-commons-java-0.2.10.jar:0.2.10]
    at com.mchange.v2.sql.SqlUtils.toSQLException(SqlUtils.java:77) ~[mchange-commons-java-0.2.10.jar:0.2.10]
    at com.mchange.v2.c3p0.impl.C3P0PooledConnectionPool.checkoutPooledConnection(C3P0PooledConnectionPool.java:690) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.c3p0.impl.AbstractPoolBackedDataSource.getConnection(AbstractPoolBackedDataSource.java:140) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at org.springframework.jdbc.datasource.lookup.AbstractRoutingDataSource.getConnection(AbstractRoutingDataSource.java:164) ~[spring-jdbc-4.3.24.RELEASE.jar:4.3.24.RELEASE]
    at org.hibernate.service.jdbc.connections.internal.DatasourceConnectionProviderImpl.getConnection(DatasourceConnectionProviderImpl.java:141) ~[hibernate-core-4.2.2.Final.jar:4.2.2.Final]
    at org.hibernate.internal.AbstractSessionImpl$NonContextualJdbcConnectionAccess.obtainConnection(AbstractSessionImpl.java:292) ~[hibernate-core-4.2.2.Final.jar:4.2.2.Final]
    at org.hibernate.engine.jdbc.internal.LogicalConnectionImpl.obtainConnection(LogicalConnectionImpl.java:214) ~[hibernate-core-4.2.2.Final.jar:4.2.2.Final]
    at org.hibernate.engine.jdbc.internal.LogicalConnectionImpl.getConnection(LogicalConnectionImpl.java:157) ~[hibernate-core-4.2.2.Final.jar:4.2.2.Final]
    at org.hibernate.internal.SessionImpl.connection(SessionImpl.java:550) ~[hibernate-core-4.2.2.Final.jar:4.2.2.Final]

Caused by: com.mchange.v2.resourcepool.TimeoutException: A client timed out while waiting to acquire a resource from com.mchange.v2.resourcepool.BasicResourcePool@43aeb5e0 -- timeout at awaitAvailable()
    at com.mchange.v2.resourcepool.BasicResourcePool.awaitAvailable(BasicResourcePool.java:1467) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.resourcepool.BasicResourcePool.prelimCheckoutResource(BasicResourcePool.java:644) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.resourcepool.BasicResourcePool.checkoutResource(BasicResourcePool.java:554) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.c3p0.impl.C3P0PooledConnectionPool.checkoutAndMarkConnectionInUse(C3P0PooledConnectionPool.java:758) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    at com.mchange.v2.c3p0.impl.C3P0PooledConnectionPool.checkoutPooledConnection(C3P0PooledConnectionPool.java:685) ~[c3p0-0.9.5.1.jar:0.9.5.1]
    ... 51 more

org.hibernate.engine.jdbc.spi.SqlExceptionHelper: SQL Error: 0, SQLState: null
org.hibernate.engine.jdbc.spi.SqlExceptionHelper: An attempt by a client to checkout a Connection has timed out

---

因此，这不是正确的解释“刷新（杀死并重新初始化）连接池”

池试图在此处“翻新”的是一个独立的

连接。它正在尝试刷新
连接
以备客户端使用。由于您已将
testConnectionOnCheckout
设置为true，并且数据库已“翻转”，因此您的ReadWriteConnectionTester将声明当前池中的每个连接无效（因为是只读的），直到池耗尽要尝试的连接或由于客户端超时而耗尽时间

您的ReadWriteConnectionTester是否返回
ConnectionTester。连接\u无效
或
ConnectionTester。数据库\u在发现
连接
无效时是否无效
？您应该希望它返回
ConnectionTester。数据库\u无效
以使其重置整个池。
只是为了澄清，在检测到RO模式时，我应该返回
ConnectionTester。数据库\u无效
，这将继续杀死并刷新整个连接池。继续以更大的重试延迟（可能是AcquisitioneTryDelay=30000）内网重试相同的数据（AcquisitioneTryAttents=0）。是否有更好的方法来实现相同的结果？如果返回
ConnectionTest.DATABASE\u无效
它将重置池。它将终止所有未完成的连接，并尝试替换
minPoolSize
。当它试图替换连接时，如果DBMS没有准备好，则可能一轮采集尝试失败。这很酷，它将在下一个客户端请求时重试，直到最终成功并能够重新组装自己。我明白了，这是有道理的。但是，如果ReadWriteConnectionTester返回的单个连接无效，它是否会终止并重新初始化连接（以保持最小池大小）？我见过在新的主数据库（RW）上出现连接的实例（mysql:show process_list；）。然而，有时随着时间的推移，我看到超时，如上图所示。明白了，我会尝试同样的方法。谢谢你的意见，史蒂夫。我看到了两者之间的性能权衡。非常感谢您的帮助连接池仍因相同的错误而失败。我在mysql数据库上监视了这个过程，几分钟后我看到
writer用户
线程不断地终止和恢复（因为db被设置为RO模式），他们在一段时间内放弃了。我没有看到新线程被终止和重新连接（show processlist），它们会一直挂起，直到被
maxIdleTime
属性搅动。假设C3PO在某个点上成功，那么可以从C3PO内部执行自动刷新操作吗。