Spring实现动态数据源切换

一、前言

    近期一项目A需实现数据同步到另一项目B数据库中,在不改变B项目的情况下,只好选择项目A中切换数据源,直接把数据写入项目B的数据库中。这种需求,在数据同步与定时任务中经常需要。

    那么问题来了,该如何解决多数据源问题呢?不光是要配置多个数据源,还得能灵活动态的切换数据源。以spring+hibernate框架项目为例(引用:http://blog.csdn.net/wangpeng047/article/details/8866239博客的图片):

    

wKioL1TxHqHQ_2EIAAA2t2DPtbc761.jpg

    单个数据源绑定给sessionFactory,再在Dao层操作,若多个数据源的话,那不是就成了下图:

    wKioL1TxH1-yOkTCAAB1EzUwMXo661.jpg

    可见,sessionFactory都写死在了Dao层,若我再添加个数据源的话,则又得添加一个sessionFactory。所以比较好的做法应该是下图:

wKioL1TxIC-gAVlAAABpRIiqH-w927.jpg

    接下来就为大家讲解下如何用spring来整合这些数据源,同样以spring+hibernate配置为例。


扫描二维码关注公众号,回复: 3572514 查看本文章

二、实现原理

    1、扩展Spring的AbstractRoutingDataSource抽象类(该类充当了DataSource的路由中介, 能有在运行时, 根据某种key值来动态切换到真正的DataSource上。)

    从AbstractRoutingDataSource的源码中:

1
public  abstract  class  AbstractRoutingDataSource  extends  AbstractDataSource  implements  InitializingBean

    我们可以看到,它继承了AbstractDataSource,而AbstractDataSource不就是javax.sql.DataSource的子类,So我们可以分析下它的getConnection方法:

1
2
3
4
5
6
7
public  Connection getConnection()  throws  SQLException {  
     return  determineTargetDataSource().getConnection();  
}  
   
public  Connection getConnection(String username, String password)  throws  SQLException {  
      return  determineTargetDataSource().getConnection(username, password);  
}

    获取连接的方法中,重点是determineTargetDataSource()方法,看源码:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
/** 
      * Retrieve the current target DataSource. Determines the 
      * {@link #determineCurrentLookupKey() current lookup key}, performs 
      * a lookup in the {@link #setTargetDataSources targetDataSources} map, 
      * falls back to the specified 
      * {@link #setDefaultTargetDataSource default target DataSource} if necessary. 
      * @see #determineCurrentLookupKey() 
      */  
     protected  DataSource determineTargetDataSource() {  
         Assert.notNull( this .resolvedDataSources,  "DataSource router not initialized" );  
         Object lookupKey = determineCurrentLookupKey();  
         DataSource dataSource =  this .resolvedDataSources.get(lookupKey);  
         if  (dataSource ==  null  && ( this .lenientFallback || lookupKey ==  null )) {  
             dataSource =  this .resolvedDefaultDataSource;  
         }  
         if  (dataSource ==  null ) {  
             throw  new  IllegalStateException( "Cannot determine target DataSource for lookup key ["  + lookupKey +  "]" );  
         }  
         return  dataSource;  
     }

    上面这段源码的重点在于determineCurrentLookupKey()方法,这是AbstractRoutingDataSource类中的一个抽象方法,而它的返回值是你所要用的数据源dataSource的key值,有了这个key值,resolvedDataSource(这是个map,由配置文件中设置好后存入的)就从中取出对应的DataSource,如果找不到,就用配置默认的数据源。

    看完源码,应该有点启发了吧,没错!你要扩展AbstractRoutingDataSource类,并重写其中的determineCurrentLookupKey()方法,来实现数据源的切换

1
2
3
4
5
6
7
8
9
10
11
12
13
14
package  com.datasource.test.util.database;
 
import  org.springframework.jdbc.datasource.lookup.AbstractRoutingDataSource;
 
/**
  * 获取数据源(依赖于spring)
  * @author linhy
  */
public  class  DynamicDataSource  extends  AbstractRoutingDataSource{
     @Override
     protected  Object determineCurrentLookupKey() {
         return  DataSourceHolder.getDataSource();
     }
}

    DataSourceHolder这个类则是我们自己封装的对数据源进行操作的类:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
package  com.datasource.test.util.database;
 
/**
  * 数据源操作
  * @author linhy
  */
public  class  DataSourceHolder {
     //线程本地环境
     private  static  final  ThreadLocal<String> dataSources =  new  ThreadLocal<String>();
     //设置数据源
     public  static  void  setDataSource(String customerType) {
         dataSources.set(customerType);
     }
     //获取数据源
     public  static  String getDataSource() {
         return  (String) dataSources.get();
     }
     //清除数据源
     public  static  void  clearDataSource() {
         dataSources.remove();
     }
 
}

    2、有人就要问,那你setDataSource这方法是要在什么时候执行呢?当然是在你需要切换数据源的时候执行啦。手动在代码中调用写死吗?这是多蠢的方法,当然要让它动态咯。所以我们可以应用spring aop来设置,把配置的数据源类型都设置成为注解标签,在service层中需要切换数据源的方法上,写上注解标签,调用相应方法切换数据源咯(就跟你设置事务一样):

1
2
@DataSource (name=DataSource.slave1)
public  List getProducts(){

    当然,注解标签的用法可能很少人用到,但它可是个好东西哦,大大的帮助了我们开发:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
package  com.datasource.test.util.database;
 
import  java.lang.annotation.*;
 
@Target ({ElementType.METHOD, ElementType.TYPE})
@Retention (RetentionPolicy.RUNTIME)
@Documented
public  @interface  DataSource {
     String name()  default  DataSource.master;
 
     public  static  String master =  "dataSource1" ;
 
     public  static  String slave1 =  "dataSource2" ;
 
     public  static  String slave2 =  "dataSource3" ;
 
}

三、配置文件

    为了精简篇幅,省略了无关本内容主题的配置。

    项目中单独分离出application-database.xml,关于数据源配置的文件。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
<? xml  version = "1.0"  encoding = "UTF-8" ?>
<!-- Spring 数据库相关配置 放在这里 -->
< beans  xmlns = "http://www.springframework.org/schema/beans"
        xmlns:xsi = "http://www.w3.org/2001/XMLSchema-instance"
        xmlns:aop = "http://www.springframework.org/schema/aop"
        xmlns:tx = "http://www.springframework.org/schema/tx"
        xsi:schemaLocation="http://www.springframework.org/schema/beans
        http://www.springframework.org/schema/beans/spring-beans.xsd
        http://www.springframework.org/schema/aop
         http://www.springframework.org/schema/aop/spring-aop-3.0.xsd
         http://www.springframework.org/schema/tx
         http://www.springframework.org/schema/tx/spring-tx-3.0.xsd">
 
     < bean  id = "dataSource1" class = "com.mysql.jdbc.jdbc2.optional.MysqlDataSource">   
         < property  name = "url"  value = "${db1.url}" />
         < property  name = "user" value = "${db1.user}"/>
         < property  name = "password" value = "${db1.pwd}"/>
         < property  name = "autoReconnect"  value = "true" />
         < property  name = "useUnicode"   value = "true" />
         < property  name = "characterEncoding"  value = "UTF-8" />
     </ bean >
 
     < bean  id = "dataSource2" class = "com.mysql.jdbc.jdbc2.optional.MysqlDataSource">
         < property  name = "url"  value = "${db2.url}" />
         < property  name = "user" value = "${db2.user}"/>
         < property  name = "password" value = "${db2.pwd}"/>
         < property  name = "autoReconnect"  value = "true" />
         < property  name = "useUnicode"   value = "true" />
         < property  name = "characterEncoding"  value = "UTF-8" />
     </ bean >
 
     < bean  id = "dataSource3" class = "com.mysql.jdbc.jdbc2.optional.MysqlDataSource">
         < property  name = "url"  value = "${db3.url}" />
         < property  name = "user" value = "${db3.user}"/>
         < property  name = "password" value = "${db3.pwd}"/>
         < property  name = "autoReconnect"  value = "true" />
         < property  name = "useUnicode"   value = "true" />
         < property  name = "characterEncoding"  value = "UTF-8" />
     </ bean >
     <!-- 配置多数据源映射关系 -->
     < bean  id = "dataSource"  class = "com.datasource.test.util.database.DynamicDataSource" >
         < property  name = "targetDataSources" >
             < map  key-type = "java.lang.String" >
         < entry  key = "dataSource1"  value-ref = "dataSource1" ></ entry >
                 < entry  key = "dataSource2"  value-ref = "dataSource2" ></ entry >
                 < entry  key = "dataSource3"  value-ref = "dataSource3" ></ entry >
             </ map >
         </ property >
     <!-- 默认目标数据源为你主库数据源 -->
         < property  name = "defaultTargetDataSource"  ref = "dataSource1" />
     </ bean >
 
     < bean  id = "sessionFactoryHibernate"  class = "org.springframework.orm.hibernate3.LocalSessionFactoryBean" >
         < property  name = "dataSource"  ref = "dataSource" />
         < property  name = "hibernateProperties" >
             < props >
                 < prop  key = "hibernate.dialect" >com.datasource.test.util.database.ExtendedMySQLDialect</ prop >
                 < prop  key = "hibernate.show_sql" >${SHOWSQL}</ prop >
                 < prop  key = "hibernate.format_sql" >${SHOWSQL}</ prop >
                 < prop  key = "query.factory_class" >org.hibernate.hql.classic.ClassicQueryTranslatorFactory</ prop >
                 < prop  key = "hibernate.connection.provider_class" >org.hibernate.connection.C3P0ConnectionProvider</ prop >
                 < prop  key = "hibernate.c3p0.max_size" >30</ prop >
                 < prop  key = "hibernate.c3p0.min_size" >5</ prop >
                 < prop  key = "hibernate.c3p0.timeout" >120</ prop >
                 < prop  key = "hibernate.c3p0.idle_test_period" >120</ prop >
                 < prop  key = "hibernate.c3p0.acquire_increment" >2</ prop >
                 < prop  key = "hibernate.c3p0.validate" >true</ prop >
                 < prop  key = "hibernate.c3p0.max_statements" >100</ prop >
             </ props >
         </ property >
     </ bean >
 
     < bean  id = "hibernateTemplate"  class = "org.springframework.orm.hibernate3.HibernateTemplate" >
         < property  name = "sessionFactory"  ref = "sessionFactoryHibernate" />
     </ bean >
 
     < bean  id = "dataSourceExchange"  class = "com.datasource.test.util.database.DataSourceExchange" />
 
     < bean  id = "transactionManager"  class = "org.springframework.orm.hibernate3.HibernateTransactionManager" >
         < property  name = "sessionFactory"  ref = "sessionFactoryHibernate" />
     </ bean >
 
     < tx:advice  id = "txAdvice"  transaction-manager = "transactionManager" >
         < tx:attributes >
             < tx:method  name = "insert*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "add*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "update*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "modify*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "edit*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "del*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "save*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "send*"  propagation = "NESTED"  rollback-for = "Exception" />
             < tx:method  name = "get*"  read-only = "true" />
             < tx:method  name = "find*"  read-only = "true" />
             < tx:method  name = "query*"  read-only = "true" />
             < tx:method  name = "search*"  read-only = "true" />
             < tx:method  name = "select*"  read-only = "true" />
             < tx:method  name = "count*"  read-only = "true" />
         </ tx:attributes >
     </ tx:advice >
 
     < aop:config >
         < aop:pointcut  id = "service"  expression = "execution(* com.datasource..*.service.*.*(..))" />
         <!-- 关键配置,切换数据源一定要比持久层代码更先执行(事务也算持久层代码) -->
         < aop:advisor  advice-ref = "txAdvice"  pointcut-ref = "service"  order = "2" />
         < aop:advisor  advice-ref = "dataSourceExchange"  pointcut-ref = "service"  order = "1" />
     </ aop:config >
 
</ beans >


四、疑问

    多数据源切换是成功了,但牵涉到事务呢?单数据源事务是ok的,但如果多数据源需要同时使用一个事务呢?这个问题有点头大,网络上有人提出用atomikos开源项目实现JTA分布式事务处理。你怎么看?

原创作品,允许转载,转载时请务必以超链接形式标明文章 原始出处 、作者信息和本声明。否则将追究法律责任。http://linhongyu.blog.51cto.com/6373370/1615895

猜你喜欢

转载自blog.csdn.net/xiaoyao880609/article/details/51242520