Determines whether Mirror Replication Agent uses replication definitions.
true
true – Enables using replication definitions.
false – Disables using replication definitions.
If the value of the use_rssd parameter is true, Mirror Replication Agent connects to the RSSD to retrieve replication definitions for the primary database whenever the resume command is invoked.
Each time it retrieves replication definitions, Mirror Replication Agent stores the information in a cache. Mirror Replication Agent uses replication definitions stored in its cache when it generates LTL commands.
If the Log Transfer Interface component encounters an operation on a database object for which it does not have a cached replication definition, Mirror Replication Agent reconnects to the RSSD to update its replication definition cache.
If a replication definition still cannot be found for the operation, the Mirror Replication Agent instance suspends operation and goes to the Admin state, unless it is in a warm standby application.
When Mirror Replication Agent is used in a Replication Server warm
standby application, it will not suspend replication
when it cannot find a replication definition, regardless of the
value of the use_rssd parameter.
Mirror Replication Agent uses information in table and function replication definitions (that is, replication definitions for individual primary database objects) stored in the RSSD to generate more efficient LTL, and thus improve throughput in the Log Transfer Interface component, and throughout the replication system.
Accessing replication definitions in the RSSD enables the Log Transfer Interface component to improve performance by:
Omitting column names in LTL.
When columns are sent in the order specified in the replication definition, column images can be sent without column names (headings), which reduces LTL overhead.
Omitting unneeded columns in LTL.
When columns are sent as specified in the replication definition, images for unchanged columns need not be sent, which reduces LTL overhead.
Sending data for each column in the datatype specified by the replication definition.
This allows data to be handled more efficiently all the way through the replication system.
Sending database object names in the same character case as defined in the replication definition.
If the value of the use_rssd parameter is false, or no replication definition exists for the table, column data is sent in the datatype defined in the primary database.
If you use owner-qualified table names for primary tables, you must:
Set the value of the use_rssd parameter to true
Specify an owner-qualified table name when you create the replication definition
Mark the table in the primary database using the owner_on clause of the sp_setreptable procedure
