Planning for Filegroups

Filegroups are a logical containers that Microsoft SQL Server uses to group tables and indexes.

You can create filegroups using the tools provided by the database vendor. Filegroup names are case-sensitive and must exactly match the names you entered when installing Landmark products.

Options for Configuring Filegroups

Lawson supports three options for configuring filegroups:

  • Use a single filegroup.

    If you don't define filegroups in Microsoft SQL Server, the database defaults to one filegroup called PRIMARY.

  • Create user-defined, unique filegroups.

  • Let Microsoft SQL Server define filegroups for the data.

Filegroup Configuration Guidelines for Lawson Data

Application consultants should work with Lawson to determine initial space requirements for filegroups. There is a spreadsheet the consultants can fill out to estimate sizes. You usually make many adjustments as you move into production and beyond.

The easiest way to allocate space is to use one large filegroup for all data. However, performance issues frequently dictate you reduce disk contention by spreading objects out to filegroups on multiple disks.

To successfully set up and implement Microsoft SQL Server, Lawson recommends a minimum of three filegroups, one for application data, one for indexes, and one for large objects. When planning your filegroups, consider these guidelines as well:

  • Separate indexes from the tables on which they are based.

  • Separate high-use, concurrent-use tables and indexes from each other.

    This is an ongoing process and varies greatly depending on the needs of your data and business processes. Application consultants can to determine which tables and indexes are used most heavily at a site.

  • Separate large object filegroups from other data and index filegroups.

Specifying Filegroups in the Database Interface Configuration File (db.cfg)

The default db.cfg database interface configuration file contains the line:

#FILEGROUPS=FALSE

If you are using only the PRIMARY filegroup, remove the # to update the database interface configuration file entry to read:

FILEGROUPS=FALSE