Managing the Replication Agent transaction log

This section describes transaction log administration and maintenance procedures for the following Sybase Replication Agent variants:

Replication Agents for Microsoft SQL Server

The Replication Agent for Microsoft SQL Server is trigger-based. To capture replicated transactions in the primary database, each trigger-based Replication Agent uses its own proprietary transaction log, which consists of objects (tables, stored procedures, and triggers) that it creates, owns, and maintains in the primary database.

The only routine maintenance required for the trigger-based Replication Agent transaction log is truncation. If not periodically truncated, the Replication Agent transaction log objects can grow to consume all of the resources allocated to the primary database.

Replication Agent for UDB

The Replication Agent for UDB uses the native DB2 log to capture replicated transactions. The Replication Agent for UDB does create some “transaction log” objects in the primary database to store its system data, but those database objects require no routine maintenance. See the Sybase Replication Agent Primary Database Guide for information on how to automatically truncate the primary database transaction log.

Replication for Oracle

The Replication Agent for Oracle uses the native Oracle log to capture replicated transactions. The transaction log objects it creates in the primary database facilitate stored procedure replication. These database objects require no routine maintenance.

Oracle log archive operations do not occur until the Replication Agent has processed all the log records in a redo log or archive log file. After it has processed, the Replication Agent issues the archive log command to the Oracle database. For information on redo log and archive log files, see the Sybase Replication Agent Primary Database Guide.

Administration tasks

The following sections describe each Replication Agent transaction log administration and maintenance task in detail: