High availability and disaster recovery

Most enterprises have requirements for documenting and testing the ability of their infrastructure to sustain loss at various levels. These levels can range from the loss of a server to the loss of the physical infrastructure of an entire data center. The topics that follow provide guidance for deploying the components in a BMC Server Automation system in highly available and easily recoverable configurations.

• Comparing high availability and disaster recovery
• Architecture summary and resource sharing
• Application Server high availability and disaster recovery
• File server high availability and disaster recovery
• Database server high availability and disaster recovery
• Reports server high availability and disaster recovery
• PXE provisioning infrastructure high availability and disaster recovery
• PXE provisioning data store server high availability and disaster recovery

Assumptions and notes

You might need to modify these guidelines based on your company's existing infrastructure and standards. These guidelines do not attempt to document every possible infrastructure and implementation for BMC Server Automation. Rather, they target the most common infrastructures in place at large enterprises.

The guidelines are directed at enterprises that require vendor-backed solutions in their IT environments. BMC is aware of the large number of possible solutions to various high-availability and disaster-recovery problems, including high-quality and freely available products for which commercial support options are limited. The products that are mentioned in these guidelines are either commercial products or are packaged with commercial products, such as operating-system vendor distributions.

The guidelines apply to BMC Server Automation version 8.0 or later. BMC assumes that you are familiar with standard high-availability and disaster-recovery practices in the IT industry, including:

• Load-balancing applications that use technologies such as the F5 Networks Big IP family of products or similarly functioning products
• Implementing traditional active and standby clustering solutions that use a virtual IP address (VIP), associated service-related file system on shared storage, service start and stop scripts, and related concepts that are provided by clustering products offered by independent software vendors, such as Veritas Cluster Server
• Data synchronization and related concepts such as those available using the Symmetrix Remote Data Facility (SRDF) family of products from EMC
• Increasing storage availability using RAID storage systems and storage controller multipathing
• Increasing network availability using network interface controller (NIC) bonding and teaming