H.2 Installing a Windows Service

The installation of a Windows service varies in complexity with the needs of the application, but usually involves the following activities:

• Selecting the user account in which the service will run.

• Registering the service and establishing its activation mode and dependencies.

• Creating one or more file system directories to contain executables, libraries, and supporting files or databases.

• Configuring those directories with appropriate permissions so that they are accessible to the user account selected for the service.

• Creating keys in the Windows registry.

• Configuring the Windows Event Log so that the service can report status and error messages.

There are many ways to perform these tasks. For example, an administrator can execute them manually, as we discuss beginning on page 1918. Another option is to write a script or program tailored to the needs of your application. Finally, you can build an installer using a developer tool such as InstallShield.

Selecting a User Account

Before installing a service, you should give careful consideration to the user account that will run the service. Unless your service has special requirements, we recommend that you use the built-in account that Windows provides specifically for this purpose. On Windows XP and Windows Server 2003, the fully-qualified name for this account is NT Authority\LocalService; in an English locale, its name is displayed as Local Service. On Windows Vista, the account name is simply Local Service.