Replicated Transactions and Procedure Identification

In a Sybase transaction replication system, the Replication Agent and Replication Server components both provide features that allow you to identify (or select) the transactions that you want to replicate. You do not need to replicate all transactions, or all data-changing operations, in the primary database.

The ability to select transactions for replication is particularly useful when you need to implement a replication system to support an application that uses some of the tables in a database, but not all of them.

By marking tables, you identify the specific tables in the primary database for which transactions are replicated. Transactions that affect the data in marked tables are referred to as replicated transactions.

By marking stored procedures, you identify (or select) the specific procedures in the primary database that are to be replicated as applied functions. When a marked procedure is invoked in the primary database, its invocation is replicated, along with its input parameter values, to the replicate database.

The ability to select procedures for replication is particularly useful when you need to implement a replication system to support an application that uses stored procedures, or when replicating a single procedure invocation is more efficient than replicating numerous, individual data-changing operations that are produced by a single procedure invocation.

• Marking and Unmarking Tables
  Individual tables to be replicated must be marked. You can mark tables explicitly using the pdb_setreptable command or automatically during ra_admin init processing when the pdb_automark_tables configuration parameter is set to true.
• Enabling and Disabling DDL Replication
  Before you enable DDL replication, you must set the ddl_username and ddl_password configuration parameters to the user name that Replication Server uses at the replicate database when executing the DDL commands. This user name must be different from the maintenance user that was configured in the Replication Server replicate connection.
• Enabling and Disabling Replication for Marked Tables
  To temporarily stop replication for a marked table (for example, when maintenance operations are performed in the primary database), you can disable replication for a marked table without affecting replication for other tables in the primary database. Then, when you are ready to resume replication from that table, you can enable replication for that table without affecting other tables in the database.
• Enabling and Disabling Replication for LOB Columns
  In this document, all columns that contain large-object (LOB) datatypes are referred to as LOB columns, regardless of the actual datatype name used by the primary database vendor. To replicate transactions that affect a LOB column, replication must be enabled for that column.
• Marking and Unmarking Stored Procedures
  Replication Agent supports Replication Server function replication by replicating the invocation of stored procedures in the primary database.
• Enabling and Disabling Replication for Stored Procedures
  To temporarily suspend replication of a stored procedure, use the pdb_setrepproc command to disable replication for the marked stored procedure. When you are ready to resume replication of the marked stored procedure, use the pdb_setrepproc command to enable replication.
• Marking and Unmarking Oracle Sequences
  Replication Agent supports replication of sequences in the primary database. To replicate a sequence invoked in a primary database, the sequence must be marked for replication and replication must be for all of that sequence. (This is analogous to marking and enabling replication for tables.)
• Enabling and Disabling Replication for Sequences
  To temporarily suspend replication of a marked sequence, you can use the pdb_setrepseq command. When you are ready to resume replication of the marked sequence, use pdb_setrepseq again to reenable replication.