Microsoft SQL Server 2000 High Availability
Authors: Hirt A.,  Cook C.,  Tripp K. L.
Published year: 2006
 Pages: 18-20/137
Buy this book on amazon.com >>

System and Process Standardization

Configuration and process management are subjects unto themselves . For purposes of this high availability discussion, the cornerstone of configuration management is to develop a standardized base configuration for all classes and types of servers. Although servers will need to be deployed that deviate from your standard, the standard is still useful because it sets a minimum expectation of performance, as well as a guide to where to find everything and what settings you can assume are in place. Anything that varies from the standard can then be documented with less effort. This is a simpler approach than trying to remember numerous server configurations individually. Similarly, administrative processes across groups can be derived to ensure a consistent, repeatable, and reliable experience.

Standardization can be accomplished through the establishment of technical specifications for products, working methods , and similar components to create system uniformity . Use of the standard specifications can be made mandatory for subordinate organizations. Ideally, you want to create standards for everything (although dont make that a laborious process in itself) and then simply reference the qualities of a system that diverge from the standard. This is much simpler than documenting each individual system separately.

Standardized administrative and support strategies, standardized directories, drive letters used, and hardware configuration make it possible to come closer to ideal manageability. The real advantage to having a standards-based system is that decisions can be made more reliably and more accurately based on familiar standardizations.



Documentation

Documentation has been mentioned briefly numerous times in both this chapter and in Chapter 1, but it needs to be mentioned again in its own right. Without good documentation, a data center will not be able to function smoothly. If the operator responsible for the day-to-day monitoring does not know the system configuration and its purpose, how can he or she know that there is problem if a blip occurs in Performance Monitor? Are there baseline performance benchmarks to compare against? Baselining and monitoring are discussed in Chapter 15, Monitoring for High Availability.

The document most data centers will have is what is referred to as a run book . A run book is a collection of documentsfrom system configurations to disaster recovery plans and everything in betweenthat will become the primary resource for system administrators and developers alike. The run book needs to be kept up to date with any changes, and all changes should stay in the run book so anyone who comes into contact with it will understand the history. Chapter 12, Disaster Recovery Techniques for Microsoft SQL Server, discusses run books in more detail, including how to create one.



Summary

Anything built without a solid foundation risks crumbling. Following the basic tenets of high availability is crucial for success; avoiding the common pitfalls is not easy. In fact, more basic tenets that are specific to certain scenarios and technologies are covered later in the book, and others mentioned here are expanded on as they relate to a specific technology or scenario. Concepts such as reliability, security, growth, and performance are all recurring themes throughout this book.

Proper staffing is crucial, and the company benefits most from DBAs who can act as leaders in the data center. Beyond people, a major part of a successful data center is how it is configured and managed, as well as the process of how changes flow from environment to environment. The systems, processes, and agreements put in place must be realistic and maintainable , and they must be documented. All parties involvedfrom executive sponsors to administrators, end users to a vendors support engineers must have similar expectations and know how the entire process works. Any failures or missed steps along the way could prove fatal.

These first two chapters lay the foundation for the rest of the book, which is largely technology-specific .


Microsoft SQL Server 2000 High Availability
Authors: Hirt A.,  Cook C.,  Tripp K. L.
Published year: 2006
 Pages: 18-20/137
Buy this book on amazon.com >>

Similar books