1. MySQL replicator that supports various target form.
2. Parallel processing on a specific column basis
3. Data change processing according to binlog event.

mvn install
java -jar target/uldra-replicator-0.0.2.jar --config-file="uldra-config.yml"

workerCount: 16
wokerQueueSize: 500

binlogServer: 127.0.0.1:3306
binlogServerID: 5
binlogServerUsername: repl
binlogServerPassword: repl
binlogInfoFile: "binlog.info"

binlogPolicies:
- name: origin.s_user
  groupKey: userid
  replicatPolicies:
  - name: "t_user"
- name: origin.s_user_status
  groupKey: userid

targetDataSource:
- !!org.apache.commons.dbcp2.BasicDataSource
  driverClassName: com.mysql.jdbc.Driver
  url: jdbc:mysql://127.0.0.1:3306/target?autoReconnect=true&useSSL=false&sessionVariables=SQL_MODE='NO_AUTO_VALUE_ON_ZERO'
  username: target
  password: target
  maxTotal: 30
  maxWaitMillis: 100
  validationQuery: SELECT 1
  testOnBorrow: false
  testOnReturn: false
  testWhileIdle: true
  timeBetweenEvictionRunsMillis: 60000
  minEvictableIdleTimeMillis : 1200000
  numTestsPerEvictionRun : 10

• name : source table name ex) {{database}}.{{table_name}}
• groupKey: data grouping column for parallel processing, included in pk is recommended.
• rowHandlerClassName: custom handler to modify row from binlog, must implement net.gywn.binlog.handler.RowHandler
• rowHandlerParams: parameters for row handler
• replicatPolicies: replicate policies for target table
  • name : target table name
  • softDelete: update null except pk if delete event income
  • upsertMode: convert insert to upsert query
  • colums: target columns to replicate, default is origin columns

• If there is more than one in targetDataSource, the data source is managed as a shardingsphere.
  • default sharding algorithm: net.gywn.binlog.shardingsphere.PreciseShardingCRC32

Currently, only a single mysql target is implemented.

Still in development. Monitoring and logging implementation is required.

[email protected]