Table or view 'undo_log' does not exist

[SilenceDevelopers]

shardingsphere-jdbc 5.5.0 org.apache.seata 2.2.0
my yaml:

spring:
  application:
    name: cola-consumer
  datasource:
    driver-class-name: org.apache.shardingsphere.driver.ShardingSphereDriver
    url:  #jdbc:shardingsphere:classpath:sharding.yaml

my sharding.yaml:

dataSources:
  ds_0:
    driverClassName: com.mysql.cj.jdbc.Driver
    dataSourceClassName: com.alibaba.druid.pool.DruidDataSource
    url: jdbc:mysql://192.168.78.201:3306/tb_order?useUnicode=true&characterEncoding=utf-8&useSSL=false
    username: root
    password: 123456
  ds_1:
    driverClassName: com.mysql.cj.jdbc.Driver
    dataSourceClassName: com.alibaba.druid.pool.DruidDataSource
    url: jdbc:mysql://192.168.78.201:3307/tb_order?useUnicode=true&characterEncoding=utf-8&useSSL=false
    username: root
    password: 123456


rules:
  - !SHARDING
    tables:
      sharding_order:
        actualDataNodes: ds.sharding_order_${0..1}
        databaseStrategy: # 分库策略，缺省表示使用默认分库策略，以下的分片策略只能选其一
          none:
        tableStrategy:
          standard:
            shardingColumn: id
            shardingAlgorithmName: table_inline
    shardingAlgorithms:
      table_inline: # 分表策略
        type: INLINE
        props:
          # strategy: STANDARD
          algorithm-expression: sharding_order_$->{id % 2 + 1}
  - !READWRITE_SPLITTING
    dataSources:
      ds:
        writeDataSourceName: ds_0
        readDataSourceNames:
          - ds_1
        loadBalancerName: round-robin
    loadBalancers:
      round-robin:
        type: RANDOM

props:
  sql-show: true

transaction:
  defaultType: BASE
  providerType: Seata

my undo_logs:

when I start up, the error is :Error creating bean with name 'dataSource' defined in class path resource [org/springframework/boot/autoconfigure/jdbc/DataSourceConfiguration$Hikari.class]: Table or view 'undo_log' does not exist.

my datasource is com.alibaba.druid.pool.DruidDataSource

I don't know why

[linghengqian]

• Can you share a git containing unit tests for testcontainers-java ? The master branch has no problems with the seata integration tests for postgres .
• By the way, the docker image of seata server does not carry the mysql jdbc driver jar starting from v2.1.0.

[SilenceDevelopers]

ok I will upload my project sooner.

[SilenceDevelopers]

my docker images:

nacos/nacos-server                  v2.4.2    325b84d58cf7   6 weeks ago     285MB
canal/canal-server                  v1.6.1    9bf4fbb74f65   7 weeks ago     1.69GB
seataio/seata-server                latest    9ba033ad828e   2 months ago    633MB
apache/rocketmq                     v5.3.0    4802c0635d11   3 months ago    419MB
mysql                               8.1.0     ae2502152260   15 months ago   574MB
redis                               5.0.14    99ee9af2b6b1   23 months ago   110MB

I upload my test project:
https://github.com/SilenceDevelopers/cola-consumer.git

[linghengqian]

@SilenceDevelopers I don't need all the relevant code, I just need a unit test execution process similar to https://github.com/linghengqian-scratches/code-repo/tree/re-test. The git you uploaded seems to contain multiple unit tests. Which unit test throws the error log you refer to? What are the steps to reproduce?

[SilenceDevelopers]

oh,sorry,I misunderstood. I didn't do a unit test. just When I was integrating the project, I found that the whole project would not start.

[SilenceDevelopers]

I found out from the logs that shardingjdbc references Hikari and Druid, but the database driver I configured did not reference Hikari.

[linghengqian]

I found out from the logs that shardingjdbc references Hikari and Druid, but the database driver I configured did not reference Hikari.

• I have explained a lot about alibaba druid in the Chinese article https://www.yuque.com/linghengqian/meve2v/zoy35sivtf6nzp90 mentioned in [Article Contribution] Welcome to Leave Your Article Link Here #12258 (comment) .
• By default, the metadata of shardingsphere is stored in the h2database in memory mode, which will create an additional hikaricp database connection pool.
• If you do not want to create a hikaricp database connection pool, consider the Cluster Mode provided by https://shardingsphere.apache.org/document/current/cn/user-manual/common-config/builtin-algorithm/metadata-repository/ , or create your own SPI implementation.
• Before Add property closeConnOnFatalError to match HikariCP's behavior alibaba/druid#6143 is reviewed, I would not recommend using alibaba druid. According to connection disabled when use com.alibaba.druid.pool.DruidDataSource #32765, druid has at least 4 bugs that existed 7 years ago that affect third-party application integration, although it does not affect simple use. This is why shardingsphere's master branch does not test druid at all.

[SilenceDevelopers]

OK,Thank you for your answer.

[linghengqian]

• By the way, if your seata server only uses file mode, you don't need to provide the mysql jdbc driver jar for the seata server's docker container. The seata server's docker container normally does not need to communicate with the mysql server unless you save the seata server's metadata in a mysql server and change the seata server's mode.