TCapture Naming Convention

• CDC Change Data Capture

• Source Node called also primary/master/producer

• Target Node called also slave/secondary/consumer

• Replication Slots types: rdb_<node>_bdr <master>_publ_<slave>

  • name rdb_<node>_bdr logical slot with plugin rdblogdec on master db
  • name <master>_publ_<slave> logical slot with plugin pgoutput on rdb_db__<master> db is the master node publication <master>_publ having a slave subscription <slave>_subs_<master>

• Publication * pubname: <master>_publ on rdb_db_<master>`

• Subscription

  • subname: <slave>_subs_<master>
  • subconninfo: primary node connection informations
  • subpublications: <node master>_publ
  • subdb: rdb_db__<node slave>

• MultiMaster MultiMaster 2 nodes MultiMaster 3 nodes MultiMaster 4 nodes

• Queue table walq_<node> structure

  • wid incremental sequence
  • lsn Logical Sequence Numner
  • xid transaction as is read from replication slot
  • data text column containing DML
  • dateop timestamp
  • current_xid transaction xid of insert into queue of dml

• Incoming/Replicated queues

• Outcoming/Replication queues

• TCapt Module to capture change on master database

• TAppl Module to apply change on slave database

• TMoni Module to monitor replication flow

• TCRepSrv TCapture Main Replication Server Program

• TC_srvctl.sh TCapture CLI

• TWrapperSQL Wrapper program to execute SQL/DML locally not replicated

• TSkipXid Skip transactions program to increase/advance replication marker on Slave

• H/M/S rispectively Monitor/Master/Slave, type of nodes

• RDB db rdb_db__<node> Replication Distribution Database

• schema: _rdb_bdr

• conf files: <node>_rdb_bdr.conf node master conf

  <node>_bdr_rdb.conf node slave conf

• tables: _rdb_bdr.

  walq_<node> queue storing replicate transactions walq_<node>_offset marker of last transaction consumed by slave walq_<node>_xid list of transactions ids incoming from replicate queues * to avoid recursive transactions (avoid cross CDC between 2 nodes with infinitive loop problem)

  xid coming from replicated queues must not be trace and insert in replication queues

• software folder structure: * path $RDBBDR * conf * sql * src * bin * run * tmp * ecc..

• features:

  • CDC stored in queues
  • transaction level granularity and commit order
  • multimaster replication in architecture of type mesh
  • automatic truncation of queues
  • table replication filtering
  • multimaster 2/3/4 nodes
  • conflict detection (not prevention)
  • history of replicated transaction stored in log tables
  • java configuration script
  • monitoring : status of master/slaves/monitor threads
  • monitoring : status of replications,positions last updated queues
  • Wrapper program to execute sql locally * not captured transaction/not in replication flow
  • multithread TCapture Replication Server
  • thread TCapt TAppl TMoni able to be started/stopped/disabled/enabled with Repl Serv in execution
  • tracing of DDL execution _rdb_bdr.walq__<node>_ddl using event trigger
  • helatcheck with upserts to move database/healthcheck
  • loglevel granularity