• A Reporting Services installation includes the report server catalog, which containsall the metadata necessary to run reporting services, including the encrypted reportdefinitions, encrypted data sources, and site and report security settings.
• The Reporting Services Configuration Manager tool helps in the setup of an SSRS instance after installation, including setup of virtual directories, service accounts, Report Server database access, and encryption keys.
• The RSConfig.exe command-line utility provides SSRS connection administration to the Report Server databases, ReportServer and ReportServerTempDB.
• The RSKeyMgmt.exe command-line utility helps manage encryption keys for backup, restore, scale-out instance management, and deletion of encrypted content.
• The RS.exe command-line utility provides a way to run SSRS script files at a command line.
• The SSRS encryption keys need to be backed up to an .snk file and stored securely to allow for a restore at a later point in the case of a sever failure or a reconfiguration of hardware.
• SSRS can be implemented in a scale-out architecture by sharing the report server catalog between multiple instances of the report server. The Reporting Services Configuration Manager tool can synchronize the encryption keys between the servers in order to access the shared catalog, or you can use the RSKeyMgmt.exe and RSConfig.exe command-line utilities to set up the scale-out environment.
• Reporting Services uses several configuration files for advanced settings and management. The main file is the RSReportServer.config file, which contains service settings, Report Manager settings, rendering device properties, delivery method settings, and so on.

Scaling out your SSRS Servers

Case scenario

As you begin the process of setting up a SQL Server 2008 BI environment, you are planning out your hardware architecture. Your solution will consist of an SSRS installation that will access online analytical processing (OLAP) data in SSAS built from a SQL Server data mart. Your manager has approved two blade servers for SSRS so that the service can be configured for scale-out deployment—that will help to balance the workload and will provide some load balancing. Additionally, your system administrator has set up Network Load Balancing (NLB) to distribute the report requests. How will you handle the following two setup requirements?

1. Your SSRS implementation requires that you set up a scale-out deployment of SSRS on the two allocated servers. You have already installed the first SSRS instance on one of the servers that points to a Report Server database on your database server. You are now ready to install an instance of SSRS on the second computer. How do you use command-line utilities to install and configure this instance to be part of the SSRS scale-out deployment?
2. To prevent a single point of failure for SSRS, you also need to handle failures at the report server catalog database level. In the case of a server hardware failure, you do not want the SSRS implementation to also fail. Which is the best high-availability technology to use for this situation and why?

1. Because the first SSRS server has been installed and configured, your next task is to install an unconfigured instance of SSRS, without having the installation use the default configuration, on the second server. You will then need to configure the virtual directories through Reporting Services Configuration Manager and set the service accounts. Before joining the new instance to the scale-out deployment, connect the SSRS instance to the same Report Server database by using RSConfig.exe with the /c parameter. At this point, you should run the RSKeyMgmt.exe tool with the /j command parameter to join an existing SSRS. You need to run this command statement on the server that is already configured and then reference the new instance that will join the existing scale-out deployment.

2. High availability at the database level can be handled by clustering, database mirroring, replication, or log shipping. However, the best choice is to use clustering. Clustering will prevent a server hardware failure from affecting the SSRS implementation, as the Database Engine will fail over to another available node in the cluster. In addition, this will be seamless to Reporting Services, as the database connection uses the virtual instance name of the database server, which also moves to the other server. With the other technologies, the SSRS implementation would need the catalog database connection to be repointed to the backup database from the mirroring, log shipping, or replication.