Spring Boot连接4个数据源失败：Not a managed type报错咨询

Hey there! Great question—Spring Boot absolutely supports connecting to more than 2 data sources; there’s no hard limit built into the framework. That error you’re seeing isn’t about the number of datasources, but rather a common configuration issue with how your entity classes are being scanned and mapped to the right datasource.

This error pops up when Spring Data JPA can’t recognize an entity class as a managed bean for a specific datasource. The most common causes are:

• You didn’t explicitly set the basePackages (or basePackageClasses) for each datasource’s LocalContainerEntityManagerFactoryBean, so some entities aren’t being scanned by the correct entity manager.
• You mixed up primary and non-primary datasource configurations, leading to missing mappings for your additional datasources.
• (Less likely, but worth checking) One of your entity classes is missing the @Entity annotation—though you probably verified this since 2 datasources already work!

Let’s walk through the key steps to resolve this:

1. Configure Each Datasource with Its Own Entity Manager and Transaction Manager
   Every datasource needs its own set of beans to avoid conflicts. Here’s an example snippet for a third datasource:

   @Configuration
   @EnableJpaRepositories(
       basePackages = "com.yourproject.repos.datasource3",
       entityManagerFactoryRef = "thirdEntityManagerFactory",
       transactionManagerRef = "thirdTransactionManager"
   )
   public class ThirdDatasourceConfig {
   
       @Bean(name = "thirdDataSource")
       @ConfigurationProperties(prefix = "spring.datasource.third")
       public DataSource thirdDataSource() {
           return DataSourceBuilder.create().build();
       }
   
       @Bean(name = "thirdEntityManagerFactory")
       public LocalContainerEntityManagerFactoryBean thirdEntityManagerFactory(
               EntityManagerFactoryBuilder builder,
               @Qualifier("thirdDataSource") DataSource dataSource) {
           return builder
                   .dataSource(dataSource)
                   .packages("com.yourproject.entities.datasource3") // Critical: point to entities for this DB
                   .persistenceUnit("thirdPU")
                   .build();
       }
   
       @Bean(name = "thirdTransactionManager")
       public PlatformTransactionManager thirdTransactionManager(
               @Qualifier("thirdEntityManagerFactory") EntityManagerFactory entityManagerFactory) {
           return new JpaTransactionManager(entityManagerFactory);
       }
   }
   

   Repeat this pattern for your fourth datasource, making sure all bean names and package paths are unique.

2. Mark Your Primary Datasource Clearly
   Your first (primary) datasource should have @Primary annotations on its DataSource, EntityManagerFactory, and TransactionManager beans. This tells Spring which one to use as the default when no specific qualifier is provided.

3. Separate Entities and Repositories by Datasource
   Organize your code into distinct packages for each datasource to avoid confusion. For example:

   • com.yourproject.entities.db1 and com.yourproject.repos.db1 for your first database
   • com.yourproject.entities.db2 and com.yourproject.repos.db2 for the second
   • And so on for db3 and db4
     This makes it easy to set the correct basePackages in each datasource’s configuration.

4. Verify Your Configuration Properties
   Ensure your application.properties (or application.yml) uses unique prefixes for each datasource, like:

   spring.datasource.first.url=jdbc:...
   spring.datasource.first.username=...
   spring.datasource.second.url=jdbc:...
   spring.datasource.second.username=...
   spring.datasource.third.url=jdbc:...
   spring.datasource.fourth.url=jdbc:...
   

Spring Boot doesn’t limit how many datasources you can connect to—you can use 4, 10, or more without issue. The error you’re facing is all about properly mapping each set of entities to its corresponding datasource’s entity manager. Follow the steps above, and you’ll have all 4 datasources up and running in no time!

内容的提问来源于stack exchange，提问作者WojteMac