TiDB
| 功能 | 说明 |
|---|---|
结构迁移 | 如目标不存在所选表，则自动根据源端元数据，结合映射生成对端创建语句并执行创建 |
全量数据迁移 | 逻辑迁移，通过顺序扫描表数据，将数据分批写入到对端数据库 |
增量实时同步 | 支持 INSERT, UPDATE, DELETE 常见 DML 同步 |
数据校验和订正 | 全量数据校验，并可选根据校验结果订正差异数据，支持定时，文档：创建定时校验订正任务 |
修改订阅 | 新增、删除、修改订阅表，支持历史数据迁移，文档：修改订阅 |
重置位点 | 按 时间戳 回溯位点，重新消费过去一段时间未被 TiKV GC 掉的增量数据 |
表名映射 | 支持 和源端保持一致, 转小写, 转大写, 以'_数字'后缀截取 |
DDL 同步 |
|
元数据检索 | 从源端表查对端，查询设置过过滤条件的 |
高级功能
| 功能 | 说明 |
|---|---|
Removal of Target Data before Full Data Migration | Remove the existing data in the Target before running the Full Data Migration, applicable for DataJobs reruning and scheduled Full Data migrations. |
Recreating Target Table | Recreate target tables before running the Full Data Migration, applicable for DataJobs reruning and scheduled Full Data migrations. |
Incremental Data Write Conflict Resolution Rule | IGNORE: Ignore primary key conflicts (skip writing), REPLACE: Replace the entire row in case of primary key conflicts. |
Handling of Zero Value for Time | Allow setting zero value for time to different data types to prevent errors when writing to the Target. |
定时全量迁移 | 文档1：创建定时全量任务 |
自定义代码 | 文档1：创建自定义代码任务 |
数据过滤条件 | 支持 WHERE 条件进行数据过滤，内容为 SQL 92 子集，文档：创建数据过滤任务 |
前置条件
| 条件 | 说明 |
|---|---|
Permissions for Account | |
Connection to PD Nodes | Make sure that BladePipe Workers can communicate with PD nodes.
|
TiKV GC Frequency | Set GC cycle to 24 hours or more in TiDB Server.
|
TiKV Historical Data Caching | Adjust the size based on task needs.
|
任务参数
| 参数名称 | 说明 |
|---|---|
printDetailLog | Print received incremental data. It is used for determining if the source database has incremental data. |
pdHost | PD node address for DataJob requests. Format: [PD_IP]:[PD_PORT], multiple PD nodes separated by , |
cdcGrpcTimeout | Timeout for gRPC channel of PD nodes to DataJob, in ms. |
cdcStubTimeout | Timeout for each stub in gRPC channel, in ms. Auto-resubscribe the stub in case of time out. |
Tips: 通用参数配置请参考 通用参数及功能
前置条件
| 条件 | 说明 |
|---|---|
Permissions for Account | INSERT, UPDATE, DELETE, and DDL permissions. |
Port Preparation | Allow the migration and sync node (Worker) to connect to the TiDB port (e.g., port 4000). |
任务参数
| 参数名称 | 说明 |
|---|---|
keyConflictStrategy | Strategy for handling primary key conflicts during write in Incremental DataTask:
|
dstWholeReplace | Convert INSERT and UPDATE operations into full row replacement in the Target. |
writeStrategy | Strategy of writing data to the Target, including:
|
Tips: 通用参数配置请参考 通用参数及功能