Configuring Microsoft SQL Server database

For a Matillion Streaming agent to capture change events from Microsoft SQL Server tables, a Microsoft SQL Server administrator with the necessary privileges must first run a query to enable CDC on the database. The administrator must then enable CDC for each table that you want the Streaming agent to capture. These steps are based on advice in the Microsoft documentation, It's recommended that you read this documentation thoroughly. This article is in the context of Data Loader's CDC feature.

CDC captures all INSERT, UPDATE, and DELETE operations committed to the tables for which CDC is enabled, once it's deployed. The connector can then record, and broadcast these events.

Prerequisites

Before the Matillion Streaming agent can be used to load data from Microsoft SQL Server, the database must be configured. To perform these steps, the following is required:

A sysadmin user must enable change data capture for the Microsoft SQL Server database (or Azure SQL Managed Instance).
- In Azure SQL Database, the db_owner role is required to enable change data capture.
The Microsoft SQL Server Agent must be running.

If the Agent isn't running already, start the Microsoft SQL Server Agent. For more information, read Start, Stop, or Pause the SQL Server Agent Service. If it's not running, you will see an error similar to the following in the server logs:

2017-01-08 15:40:24,596 @ -ERROR cached5
com.webaction.source.tm.MSSqlTransactionManager.getStartPosition
(MSSqlTransactionManager.java:389) 2522 :
Could not position at EOF, its equivalent LSN is NULL

Enabling CDC on the SQL Server database

As noted in Prerequisites, a privileged user must first activate the database for change data capture, before a capture instance for specific tables can be established.

To determine if a database already has CDC enabled, run the following command:
```
SELECT name, is_cdc_enabled
FROM sys.databases
WHERE name = '<DATABASE_NAME>'
```
If the value for is_cdc_enabled is 1, then CDC is enabled for the database, otherwise it's not enabled.
To enable CDC on your database, run the following stored procedure:
```
USE <database name>
EXEC sys.sp_cdc_enable_db
```

Enabling the database for CDC will trigger the creation of several database-specific resources: A schema with the name cdc, a cdc user, metadata tables, and other system objects.

Creating a CDC user role

Create a SQL Server user to be used by the Microsoft SQL Server Reader. The user should have sysadmin, a fixed server role for the SQL Server, and must use the SQL Server authentication mode, which must be enabled in SQL Server.
Grant the Microsoft SQL Server Reader user the db_owner role. This must be done for each database that you want to read. Use the following commands, substituting your own details:
```
USE <database name>
EXEC sp_addrolemember @rolename=db_owner, @membername=<user name>
```

For example, to enable CDC on the database mydb, create a user cdcuser, and give that user the db_owner role on mydb:

USE mydb
EXEC sys.sp_cdc_enable_db
CREATE LOGIN cdcuser WITH PASSWORD = 'passwd'
CREATE USER cdcuser FOR LOGIN cdc
EXEC sp_addrolemember @rolename=db_owner, @membername=cdcuser

Enabling CDC on a SQL Server table

Once your database has been enabled for CDC, you must then enable CDC for all tables that you wish to capture data from. A capture instance for each source table must be created. To do so, the creating member must have the db_owner fixed database role. The stored procedure sys.sp_cdc_enable_table can then be used as described below.

Use the following command to retrieve all the source tables that have been enabled for CDC:

SELECT name, is_tracked_by_cdc, SCHEMA_NAME(schema_id) AS 'schema_name'
FROM sys.tables
WHERE is_tracked_by_cdc = 1

Use the following command to enable CDC on source tables you want to capture:

USE MyDB
GO
EXEC sys.sp_cdc_enable_table
@source_schema = N'dbo',
@source_name = N'MyTable',
@role_name = N'MyRole',
@filegroup_name = N'MyDB_CT',
@supports_net_changes = 0
GO

The following fields are used in the above script:

Field name	Description
@source_name	Specifies the name of the table that you want to capture.
@role_name	Specifies a role you can add users to, who you want to grant `SELECT` permission on the captured columns of the source table. `sysadmin` or `db_owner` roles also have access to the specified change tables. Set the value of `@role_name` to `NULL`, to allow only members in the `sysadmin` or `db_owner` roles to have full access to captured information.
@filegroup_name	Specifies the `filegroup` where SQL Server places the change table for the captured table. The named `filegroup` must already exist. It's best not to locate change tables in the same `filegroup` that you use for source tables.
@supports_net_changes	For more information, read Microsoft @supports_net_changes.

Note

When @supports_net_changes is set to 1, an additional non-clustered index is created on the change table, and the net changes query function is created. Because this index needs to be maintained, enabling net changes can have a negative effect on CDC performance.

Verify user access to the CDC table

To retrieve CDC configuration information from a database or table, an SQL Server administrator can use a system stored procedure. This type of administrator with the necessary privileges must do the following:

Make sure all the captured columns in the capture instance have SELECT permissions. Members of the db_owner database role have access to all the defined capture instances' information.
You have membership in any gating roles that are defined for the table information that the query includes.
Run the sys.sp_cdc_help_change_data_capture stored procedure to query the table.

Queries shouldn't return empty results.

The following example runs the stored procedure sys.sp_cdc_help_change_data_capture on the database MyDB to query a table for CDC configuration information:

USE MyDB;
EXEC sys.sp_cdc_help_change_data_capture

The query returns configuration information for each CDC-enabled table in the database that includes change data that the caller is authorized to see. If the result is empty, make sure the user has access to the capture instance as well as the CDC tables.

SQL Server always on

The SQL Server connector can capture changes from an Always On read-only replica.

Prerequisites

Change data capture is configured and enabled on the primary node. SQL Server doesn't support CDC directly on replicas.
The read-only replica must be configured with Readable Secondary as either Read-intent only or Yes. For more information, read Microsoft's documentation, including the detailed steps in the article.
Ensure the Read Replica toggle is on when creating your pipeline.

Schema drift

Schema drift occurs when your data source changes its metadata, by adding, removing, or changing fields, columns, and types. When not handled correctly, schema drift will cause a difference between the data source, and your CDC pipeline configuration, resulting in the pipeline failing or processing incorrect information.

Due to SQL Server limitations, Data Loader is not able to automatically handle schema drift on source tables in Microsoft SQL Server databases. A SQL Server connector with CDC enabled, propagates column type changes to your destination, but it does not propagate any other schema changes, such as renamed columns, new tables, or new columns, meaning you may have to reconfigure your pipeline in response to changes in the source metadata.

To accommodate offline or online schema updates, follow the steps outlined in the following Debezium links to ensure that the changes will be reflected in your CDC pipeline:

Limitations

Change Data Capture has the following limitations:

To build a change capture instance in SQL Server, the base object must be a table. As a result, SQL Server doesn't enable collecting changes from indexed views (also known as materialized views).
If the Microsoft SQL Server Agent service isn't running, Microsoft SQL Server CDC (Change Data Capture) capture jobs won't execute.
CDC doesn't support the values for computed columns even if the computed column is defined as persisted. Computed columns that are included in a capture instance always have a value of NULL.

For additional limitations, read the Microsoft documentation.

Administration and monitoring of change capture tables

Each user setup will have nuances, so it's critical that cleanup and retention is agreed with your SQL server administrator or database administrator, and is for your use case. Appropriate monitoring should also be configured accordingly. For more information, read Administer and monitor change data capture - SQL Server.

The Debezium documentation, Debezium connector for SQL Server, documents how change data capture can impact your server, and what actions you may be able to take to help mitigate this issue.