Database clustering and storage can provide redundancy and improve high availability for the database serving the Learn installation.
Windows SQL Server Database Clustering and Storage
Setting up SQL Server Failover and setting up a separate storage device for the data files and log files are optional steps. Many institutions will not need to set up SQL Server Failover or to set up the data files and log files on a separate storage device.
Setting Up SQL Server Failover
Blackboard Learn can be run using database clustering for redundancy. Database clustering instantly transfers database requests to an identical server if the database server fails.
Blackboard Learn is designed to work with Microsoft SQL Server Failover clusters. SQL Server Failover provides added fault tolerance to the database by using another database in the cluster in the case of an error that causes the production database to fail.
A SQL Server Failover cluster is based on the same design as a Windows cluster. A group of servers are joined to provide continual service in the event of an outage or failure. When SQL Server is forced to shut down, another server steps in to handle requests.
The process for setting up a Microsoft SQL Server failover cluster is done at the operating system and database level before installing Blackboard Learn.
- Set up the Microsoft server cluster.
- Install SQL Server and configure the failover cluster.
- Set up optional file system storage
- Install Blackboard Learn.
The following is a list of tips for setting up SQL Server failover.
- Use four public IP addresses: node 1, node 2, Windows cluster server, and virtual SQL server.
- The local disk of each node must be attached to a separate SCSI controller. It should not use the same one that is attached to the RAID system.
- The RAID SCSI controller IDs of the two nodes must be different.
- RAID SCSI BIOS must be set up to enable cluster (shared bus).
- The names of the network cards cannot contain spaces or special characters.
- Do not use the default instance name for installing SQL server on a virtual server. The instance name must not include any underscores, spaces, or punctuation. Use only alphanumeric characters.
- MSDTC service must be running before starting SQL server setup.
To learn more about installing and configuring a Microsoft SQL Server failover cluster, see Microsoft TechNet at http://www.microsoft.com/technet/. While Blackboard supports SQL Server failover with Blackboard Learn, Microsoft is the best resource for information about installing and configuring a Microsoft SQL Server failover cluster.
Setting Up a Separate Storage Device for the Data Files and Log Files
The Installer will prompt for the directory where the database data files and log files should be stored. Consult the Microsoft SQL Server documentation for information about storing these files on a separate device. Keep in mind that the storage device should have a dedicated, high-bandwidth connection to the database server.
Oracle Database Clustering and Storage
These steps are optional. Many institutions will have no need to set up RAC or to set up the data files and index files on a separate storage device.
If these options are unnecessary, proceed to the instructions for setting up a separate file system server.
It is strongly recommended by Blackboard that for purposes of system management Blackboard Learn be installed on a single instance of Oracle. If for reasons of Institutional Policy you must run Blackboard Learn in a RAC environment you may do so, but Blackboard will only support the installation at the Learn-Oracle JDBC layer - not at the Oracle RAC layer. For information about running Blackboard Learn with RAC, see Oracle RAC Support.
The thin JDBC driver should be used to connect to an Oracle RAC system. However, should additional functionality be required, for example, control over network layer encryption, switching to the OCI native driver is possible.
Setup a Separate Storage Device for the Data Files and Log Files
The Installer will prompt for the directory where the database data files and log files should be stored. For information about storing these files on a separate device, see the Oracle documentation. Keep in mind that the storage device should have a dedicated, high-bandwidth connection to the database server.