Non-Sybase data servers can be either the source or the target of replication data (that is, primary or replicate databases). To move data from a non-Sybase data server requires a Replication Agent. To move data to a non-Sybase data server requires a DirectConnect database gateway.
To manage a non-Sybase data server in a replication environment, RSM must retrieve the server’s metadata, such as table names, column properties, stored procedure parameters, and so forth. Since RSM cannot directly access a non-Sybase data server, it obtains metadata either through a Replication Agent (for primary data servers), or through a DirectConnect database gateway (for replicate data servers).
In a Sybase-only replication system (with only Adaptive Server database servers), Replication Server communicates directly with all of the data servers using Open Client connectivity. Replication Server database connections are configured as either replicate-only, or primary-and-replicate (bidirectional) in a Sybase-only replication system.
In a replication system with non-Sybase data servers, a primary-only connection is used for a non-Sybase data server that is only a source of data (that is, no data flows out to the non-Sybase data server).
In a primary-only configuration, RSM retrieves the metadata for the non-Sybase data server through the Sybase Replication Agent.
The primary-only solution is implemented in RSM. The Replication Server database connection is set up as a normal primary-and-replicate connection, but RSM filters out the status of the DSI thread and reports only the state of the Rep Agent User thread. The DSI thread still exists in the Replication Server in a suspended state.
You cannot use a primary-only configuration to replicate
data from DB2 for OS/390. RSM does not manage the Replication
Agent for DB2 UDB for OS/390, so the Replication Agent
cannot retrieve the DB2 metadata for RSM. For instructions on setting
up a DB2 for OS/390 database as a replication source (primary
database), see “DB2 Universal Database for OS/390 replication environment”.
The following servers are required to set up a primary-only configuration:
Non-Sybase primary data server
Sybase Replication Agent designed for the non-Sybase database
Replication Server
Adaptive Server replicate data server
RSM Server
RSM does not support either a ping cycle or a heartbeat process for a primary-only connection because the DSI thread in the Replication Server is in a suspended state. If either the ping cycle or the heartbeat process is needed, you must set up a bidirectional (primary-and-replicate) configuration.
A replicate-only database connection is used when a non-Sybase data server is only the target of replicated data (that is, no data flows from the non-Sybase data server). The replicate-only connection is a standard Replication Server replicate database connection. For a replicate-only connection, the Replication Server contains a DSI thread, but no Rep Agent User thread. For a non-Sybase data server, replicate data passes through a DirectConnect database gateway server. RSM uses the DirectConnect server to retrieve metadata from the non-Sybase replicate data server.
All non-Sybase data servers that are supported by RSM use the same configuration to set up a replicate-only connection.
The following servers are required to set up a replicate-only configuration:
Adaptive Server primary data server
Replication Server
DirectConnect database gateway server designed for the non-Sybase replicate data server (or MainframeConnect in a gatewayless environment)
Non-Sybase replicate data server
RSM Server
The following limitations apply to replicate-only connections:
RSM does not display table or stored procedure metadata for either Adaptive Server or non-Sybase replicate databases.
RSM does not validate Maintenance User permissions at the replicate database when it creates a subscription for a non-Sybase replicate data server.
RSM does not validate the table columns or stored procedure parameters by comparing the primary and replicate databases when it creates a subscription for a non-Sybase replicate data server.
A bidirectional (or primary-and-replicate) connection is used when a non-Sybase data server acts as both the source and target of replication. For a bidirectional connection, the Replication Server database connection contains both a DSI thread and a Rep Agent User thread. RSM reports the status of both of these threads.
The bidirectional configuration uses a Sybase Replication Agent to retrieve the data from a non-Sybase data server and a DirectConnect database gateway server (or MainframeConnect in a gatewayless environment) to send data to the non-Sybase replicate database. As with the primary-only configuration, RSM uses the Sybase Replication Agent to retrieve metadata from the non-Sybase primary data server.
You cannot use this form of a bidirectional connection
to replicate data to or from a DB2 for OS/390 database.
For instructions on setting up a bidirectional DB2 for OS/390
environment, see “DB2 Universal Database for OS/390 replication environment”.
The following servers are required to set up a bidirectional configuration:
Non-Sybase data server that acts as both a primary database and a replicate database
Sybase Replication Agent designed for the non-Sybase database
Replication Server
DirectConnect database gateway server designed for the non-Sybase data server (or MainframeConnect in a gatewayless environment)
RSM Server
The following limitations apply to bidirectional connections:
RSM does not validate Maintenance User permissions at the replicate database when it creates a subscription for a non-Sybase replicate data server.
RSM does not validate the table columns or stored procedure parameters by comparing the primary and replicate databases when it creates a subscription for a non-Sybase replicate data server.
