Skip to content

MySQL Ingestion Guide

Prerequisites

CategoryDescription
VersionVersions 5.6, 5.7, and 8.x.
Settings- The source database must have ROW binlog enabled.
- It is recommended to retain the Binlog for at least 3 days.
PermissionsThe source database account requires the following minimum permissions:
SELECT, SHOW DATABASES, REPLICATION SLAVE, REPLICATION CLIENT.
Reference command: GRANT SELECT, SHOW DATABASES, REPLICATION SLAVE, REPLICATION CLIENT ON *.* TO 'username'@'%';
Tables- Tables must have a primary key.
- Primary keys for sharded tables must be consistent.

Synchronization Details

  1. The source database {db} corresponds to the data warehouse database input_{db}. The warehouse database is automatically created after the ingestion task is set up.
  2. The source table {tableName} retains the same name in the data warehouse. The table is automatically created after the ingestion task is set up.
  3. For sharded tables, the corresponding database and table names in the data warehouse are specified by the user.
  4. The synchronization impact of source database operations is as follows:
Source OperationAuto-synced to Warehouse?Impact on WarehouseRisk Level
DML data changes✅ YesData is synchronizedLow (Physical deletes are directly synchronized as deletes)
CREATE TABLE❌ NoNo changeLow (Requires manually modifying the ingestion task and backfilling data for the new table)
ADD COLUMN✅ YesThe new column is synchronizedLow (Default values for historical data in the column are not synchronized)
DROP COLUMN❌ NoNo changeLow
RENAME COLUMN⚠️ Not directly synchronizedA new column is added to receive subsequent data
MODIFY COLUMN (Increase length)✅ YesColumn length is changedLow
MODIFY COLUMN (Decrease length)❌ NoNo changeLow
MODIFY COLUMN_TYPE⚠️ Not directly synchronizedThe data ingestion task will report an error. The platform handles this through a scheduled schema synchronization task.
DROP TABLE❌ NoNo changeLow
TRUNCATE TABLE❌ NoNo changeLow

Yaoqing AI Big Data