eplication agents are the basic components of SQL Server 2000 replication. Before you settle on the type of replication you want in your application architecture, it is important to understand how each type of replication works and what agents they use. Since the different types of replication use different combinations of replication agents, understanding what those agents do helps clarify the resulting SQL Agent jobs when you install replication. For example, snapshot, transactional, and merge replication all use the Snapshot Agent to initialize the subscriber's data. Then they employ other agents to implement the replication process. In this article, you'll learn about the roles of the replication agents and learn how to inspect the properties and profiles of the Snapshot and other agents.
Replication Basics: a Review
As you learned in the previous Replication 101 article, ( "Terminology, Types, and Configuration", CoDe Magazine, Nov/Dec 2003)
, SQL Server 2000 replication uses a publish/subscribe metaphor. A publisher server contains data that will be distributed to subscribers. The specification of the data you want to replicate is called a publication
, and a publisher server may contain many publications. Each publication consists of one or more articles
, which are the database objects including tables, views, stored procedures, and user-defined functions that the publisher server will publish to the subscribers.
A subscriber is usually another server that receives data from the publisher and optionally sends updates back to it. You create a subscription to the publication using replication, which SQL Server implements using SQL Agent jobs that execute on a scheduled basis. A distributor is another server that you may use to distribute data.
|A push subscription executes a SQL Agent job on the distributor or publisher and sends the data to the subscriber, while a pull subscription executes a SQL Agent job on the subscriber to retrieve the data.|
A push subscription
executes a SQL Agent job on the distributor or publisher and sends the data to the subscriber, while a pull subscription
executes a SQL Agent job on the subscriber to retrieve the data.
When you configure replication, you define which servers will be the publisher and distributor, and optionally, which servers will be the subscribers.
After you have configured replication, you define a publication on the publisher server that consists of articles of data. When you define a publication, you must choose what type of replication the publication has: snapshot, transactional, or merge. Once you've defined the publication, you can set up one or more subscriptions to the publication from one or more servers. The steps you take in defining a publication vary depending on the type of replication you choose.