Spring Batch 不使用自定义数据源来创建表 [英] Spring Batch doesn't use custom datasource to create tables

问题描述

我正在研究 REST 服务(使用 Spring Boot)，它运行批处理作业.我希望 Batch 只与嵌入式数据源(用于存储元数据)，而默认数据源(在我的示例中为 Postgres)将用于存储业务实体.

I am working on REST-service (with Spring boot), which runs batch job. I want Batch to work only with embedded datasource (to store metadata), while default datasource (Postgres in my case) will be used to store buisness entities.

问题是 Batch 尝试在启动时在默认数据源中创建元数据表(如 batch_job_execution、batch_job_instance 等).

The problem is that Batch tries to create metadata tables (like batch_job_execution, batch_job_instance, etc.) in default datasource at startup.

以下是重现问题的示例配置:

Here is the sample configuration, which reproduces the problem:

批量配置

@Configuration
@EnableBatchProcessing
public class BatchConfiguration extends DefaultBatchConfigurer {

    @Override
    @Autowired
    public void setDataSource(@Qualifier("batchDataSource") DataSource dataSource) {
        super.setDataSource(dataSource);
    }
}

数据源配置

@Configuration
public class DataSourceConfiguration {
    @Bean
    @Primary
    public DataSource DataSource() {
        final SimpleDriverDataSource dataSource = new SimpleDriverDataSource();
        dataSource.setDriverClass(org.postgresql.Driver.class);
        dataSource.setUrl("jdbc:postgresql://localhost:5432/test_batch");
        dataSource.setUsername("user");
        dataSource.setPassword("password");

        return dataSource;
    }

    @Bean(name = "batchDataSource")
    public DataSource batchDataSource() {
        return new EmbeddedDatabaseBuilder().setType(EmbeddedDatabaseType.H2).build();
    }
}

使用此配置，我在微服务启动时在 Postgres 中获取批处理表，尽管之后似乎使用了嵌入式数据源，因为我在尝试时遇到 找不到表" H2 错误启动作业.

With this configuration I am getting batch tables in Postgres when microservice started, though aftewards, it seems, that embeded datasource is used, because I am getting "Table not found" error for H2 when trying to launch a job.

那么我应该如何正确编写配置以使 Batch 仅适用于嵌入式数据源?我不想要主数据源中的任何元数据(甚至是空表).

So how should I properly write configuration to make Batch work only with embedded datasource? I don't want any metadata (even empty tables) in main datasource.

更新:

正如 Michael Minella 所说，应该再添加一颗豆子:

As Michael Minella said, one more bean should be added:

@Configuration
@EnableBatchProcessing
public class BatchConfiguration extends DefaultBatchConfigurer {

    @Override
    @Autowired
    public void setDataSource(@Qualifier("batchDataSource") DataSource dataSource) {
        super.setDataSource(dataSource);
    }

    @Bean
    public BatchDatabaseInitializer   batchDatabaseInitializer(@Qualifier("batchDataSource") DataSource dataSource, ResourceLoader resourceLoader){
        BatchDatabaseInitializer batchDatabaseInitializer = new     BatchDatabaseInitializer(dataSource, resourceLoader, new BatchProperties());
        return batchDatabaseInitializer;
    }
}

推荐答案

使用 Spring Boot 不幸的是，BatchDataSourceInitializer 使用的 DataSource 与BatchConfigurer.它只是在上下文中获取默认的 DataSource.如果你配置你自己的BatchDataSourceInitializer，Boot 不会启动，你可以定义直接使用哪个DataSource.

Using Spring Boot the DataSource that is used by the BatchDataSourceInitializer is, unfortunately, not related to the BatchConfigurer. It just grabs the default DataSource in the context. If you configure your own BatchDataSourceInitializer, the Boot one won't kick in and you can define which DataSource is used directly.

这篇关于Spring Batch 不使用自定义数据源来创建表的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！