如何在数据库失败时运行spring引导应用程序

Question

我有一个弹簧引导应用程序，它有多个数据库。无论数据库的状态如何，我都想运行这个应用程序。下面是我的应用程序yaml的样子

logging:
  level:
    org:
      hibernate:
        SQL: DEBUG
        type:
          descriptor:
            sql:
              BasicBinder: TRACE

spring:
  sql:
    init:
      continue-on-error: true # <-- didn't work
  jpa:
    properties:
      hibernate:
        format_sql: 'true'
    show-sql: 'true'
  datasource:
    azure-read-only:
      url: <azure_db1_url>
      username: '<username>'
      password: '<password>'
      driver-class-name: com.microsoft.sqlserver.jdbc.SQLServerDriver
      continue-on-error: true
      hikari: # <--- add this to bypass boot failure but it didn't work either
        minimum-idle: 0
        maximum-pool-size: 15
        connection-timeout: 10000 #10s
        idle-timeout: 300000 #5m
        max-lifetime: 600000 #10m
        initialization-fail-timeout: -1
        validation-timeout: 1000 #1s
    azure-read-write:
      url: <azure_db2_url>
      username: '<username>'
      password: '<password>'
      driver-class-name: com.microsoft.sqlserver.jdbc.SQLServerDriver
      continue-on-error: true
      hikari:
        minimum-idle: 0
        maximum-pool-size: 15
        connection-timeout: 10000 #10s
        idle-timeout: 300000 #5m
        max-lifetime: 600000 #10m
        initialization-fail-timeout: -1
        validation-timeout: 1000 #1s
    db2-testdb1:
      url: <another_db_not_hosted_on_azure>
      username: '<username>'
      password: '<passwor'
      driver-class-name: com.ibm.db2.jcc.DB2Driver
      continue-on-error: true
      hikari:
        minimum-idle: 0
        maximum-pool-size: 15
        connection-timeout: 10000 #10s
        idle-timeout: 300000 #5m
        max-lifetime: 600000 #10m
        initialization-fail-timeout: -1
        validation-timeout: 1000 #1s

对于数据源，我有一个数据源，即Primary bean，而其他数据源是非主数据源。我怎样才能运行spring引导应用程序，即使它无法连接到所有dbs。

Answer 1

您可以“手动”实例化数据源，然后将其提供给Spring。例如：

@Configuration
public class MyConfigBean {

  @Lazy
  @Autowired
  private DataSource dataSource;
  
  @Bean
  public DataSource getDataSource() {
    log.debug(">>> Building the DATA-SOURCE");
    DataSourceBuilder<?> dataSourceBuilder = DataSourceBuilder.create();
    dataSourceBuilder.driverClassName(JDBC_DRIVER_CLASS_NAME);
    dataSourceBuilder.url(props.getJDBCUrl());
    dataSourceBuilder.username(props.getJDBCUsername());
    dataSourceBuilder.password(props.getJDBCPassword());
    return dataSourceBuilder.build();
  }
  
}

@Lazy注释将对getDataSource()方法的调用保留到最后。通过这种方式，您可以在更早的时候检索/填充URL和凭据，因此它们在需要的时候已经准备好了。