High availability for database servers

This topic provides guidelines for implementing high availability for database servers in a TrueSight Server Automation infrastructure. The following databases are addressed:

Oracle database

The Oracle solution for enterprise-level highly available databases is the Real Application Cluster (RAC) technology. This technology allows a horizontally scalable database with all nodes in an active state. Shared disks and private cluster network interconnects ensure data consistency across all members of the cluster. The solution provides a combination of failover capability and load balancing between the nodes on the cluster.

The following figure summarizes the architecture in an Oracle RAC solution:

worddavf183d590887105fbeea3b3d1bcf1cbc7.png

Oracle RAC database implementation is beyond the scope of this documentation. The RAC framework offers flexibility in its architecture. As a result, the specific implementation depends on the existing IT infrastructure in your enterprise.

BMC makes the following recommendations and assumptions regarding an Oracle RAC used as the TrueSight Server Automation database at the primary site:

  • BMC recommends that you use a symmetric configuration with no affinity between Application Servers and database nodes.
  • BMC assumes that your implementation is a multinode Oracle RAC cluster that conforms to the Oracle Maximum Availability Architecture (MAA) best practices. For information, see the following Oracle website: Oracle Maximum Availability Architecture.
  • The cluster implementation uses N database-specific virtual IP addresses (VIPs), with each VIP using a specific database node as its home. The VIPs fail over to other nodes in the cluster when their home node becomes unavailable.
  • The database-specific VIPs and the associated TCP network port for accessing the database is available to all components of the TrueSight Server Automation architecture that require a database, including the Application Servers (including provisioning servers) and report servers.

Microsoft SQL Server

The recommended technology for implementing high availability in a SQL Server database environment is the SQL Server failover cluster. For optimal and uniform hardware utilization, BMC recommends active-active cluster configuration. However, an active-passive implementation is also viable and is less costly.

SQL Server failover cluster implementation is beyond the scope of this documentation. For information about high-availability options in SQL Server, see the following web page: http://msdn.microsoft.com/en-us/library/ms190202.aspx.

Notes

  • The SQL Server cluster publishes an instance name that is referred to in the TrueSight Server Automation database connection configuration.
  • To the application, the failover cluster appears as a single database instance.
  • In an active-active cluster, the application load is evenly distributed among the nodes of the database cluster. In an active-passive cluster, only the active node services the application.
  • A failover event is transparent to the application.

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*