At the time of writing this book, there is already a fair amount of capabilities for building some highly available application solutions using Linux on the mainframe. This chapter has laid out some of the key areas where implementation is both feasible and cost-effective. In the dynamic world of Linux on the mainframe, there is still much going on.
The Open Source community has two major projects dedicated to high availability. This chapter has already discussed the LVS approach from the Linux Virtual Server project. The other key project is the High-Availability Linux project that is building key infrastructure for peer cluster environments. IBM, through its Linux Technology Center, is a very active participant in this latter project, as are some other companies.
The area of high availability has attracted significant attention from both small and large companies, as you can easily see from a Web search such as: Linux + "high availability." Each of the major Linux distributors is also addressing aspects of high availability as part of its mainframe offering.
The IBM Tivoli System Automation (ITSA) for Linux provides the high-level tools that make it much easier to build high-availability clusters from Linux images. These tools are particularly helpful for a cluster built on z/VM virtualization. ITSA for Linux uses the abstractions of: resource types, resource managers, grouping services, and automation policies. There is one resource manager for one or more resource types. Resource types are, for example, network adapters or file systems and disks. ITSA for Linux monitors changes to the resources contained in resource groups, and provides complex proactive and reactive management beyond simple failover capabilities. It moves beyond simple scripts to create policies that can be implemented throughout a cluster. These policies can be used to implement processes in which complex environments are described and automated.