BladePipe 1.9.0: New data pipelines, faster Oracle writes, and improved stability.
跳到主要内容

TiDB 到 Kafka

选择对端数据库:

数据链路

基本功能

功能说明
Schema Migration

If the specified Topic after mapping does not exist in the Target, BladePipe will automatically create the Topic, allowing setting the number of partitions.

Full Data Migration

Migrate data by sequentially scanning data in tables and writing it in batches to the target database.

Incremental Data Sync

Sync of common DML like INSERT, UPDATE, DELETE is supported.

Subscription Modification

Add, delete, or modify the subscribed tables with support for historical data migration. For more information, see Modify Subscription.

Position Resetting

Reset positions by timestamp to consume again the incremental data that has not been collected as garbage by TiKV in a past period.

Metadata Retrieval

Retrieve the target metadata with filtering conditions or target primary keys set from the source table.

高级功能

功能说明
消息格式

支持以下消息格式,文档:消息格式说明

  • CloudCanal内置格式
  • AlibabaCanal兼容格式
Topic Mapping Rules

By default, Topic is formed by connecting source instance id, database, and table with . in between (e.g., my-vgpq6q097174t6t.dingtax.app_key). Also, it supports the mapping rules, namely, keeping the name the same as that in Source, converting the text to lowercase, converting the text to uppercase.

Table-level Topic

Create Topics corresponding to the tables in the Source, and the table partitions can be obtained automatically.

DDL Dedicated Topic

Allow specifying a Topic for DDL. If not specified, DDL time is placed in partition 0 of the Topic created from the corresponding table.

Scheduled Full Data Migration

For more information, see Create Scheduled Full Data DataJob.

Custom Code

For more information, see Custom Code Processing, Debug Custom Code and Logging in Custom Code.

Data Filtering Conditions

Support data filtering using WHERE conditions, with SQL-92 as the SQL language. For more information, see Data Filtering.

使用示例

标题详情
跨互联网数据互通 (Kafka)

文档:跨互联网数据互通 (Kafka)

Kafka 数据中转校验

文档:Kafka 数据中转校验


源端数据源

前置条件

条件说明
账号权限

文档:TiDB 需要的权限

PD节点网络连通

请确保 CloudCanal 各节点能正常与 PD 各节点通讯

  • telnet [PD节点IP] [PD节点端口号]
TiKV GC 回收频率

在 TiDB Server 中修改 GC 周期时间为 24小时 以上

  • set global tidb_gc_life_time = "24h0m0s";
TiKV 历史变更数据缓存

建议根据任务所需适当调整大小

  • old-value-cache-memory-quota:增量旧数据占用 TiKV 节点的内存的上限
  • sink-memory-quota:增量数据占用 TiKV 节点的内存的上限

任务参数

参数名称说明
printDetailLog

打印接收到的增量,常用于判断源端是否有增量数据推送

pdHost

任务请求的 PD 节点地址,格式为: [PD_IP]:[PD_PORT], 多个 PD 节点用 , 隔开
例: 127.0.0.1:2379,127.0.0.1:2380

cdcGrpcTimeout

任务与 PD 节点 gRpc 连接通道的超时时间,单位ms

cdcStubTimeout

gRpc 通道中的每个 stub 的超时时间,超过该时间会自动重新订阅,单位ms

fastFailKeywords

字符串数组,以逗号分隔,当异常信息中包含这些关键字时,任务不再尝试重连,直接重启。例如 DEADLINE_EXCEEDED 表示当 gRPC 超时异常时不再重连,直接重启任务

Tips: 通用参数配置请参考 通用参数及功能


目标端数据源

前置条件

条件说明
网络准备

迁移同步节点(sidecar)可连接 Kafka 各节点

任务参数

参数名称说明
schemaFormat

消息格式,文档:消息格式说明

batchWriteSize

单条消息最大数据条数,超过则拆分消息

defaultTopic

无法找到对应 Topic 的消息则发送到此 Topic (如新增表)

ddlTopic

专门发送 DDL 的 Topic, 为空则发送到对应 Topic 的第 0 个分区

compressionType

Kafka compression.type 参数, 设置压缩算法, 支持 GZIP, SNAPPY, LZ4, ZSTD 算法

batchSize

Kafka batch.size 参数

acks

Kafka acks 参数, 默认 all

maxRequestBytes

Kafka max.request.size 参数

lingerMs

Kafka linger.ms 参数, 默认 1

envelopSchemaInclude

当 schemaFormat 设置为 DEBEZIUM_ENVELOP_JSON_FOR_MQ 时,消息体是否包含 schema 信息

customClientProps

自定义传入到 Kafka Client 参数,JSON 格式,key为参数名,value为参数值。此配置项以最高优先级生效。例如:AWS IAM 访问控制

Tips: 通用参数配置请参考 通用参数及功能