New DB2 WLM Best Practice Paper
svisser1 2700018UK9 Visits (2614)
Congratulations to the Best Practices team for putting out another excellent paper. This one is for Workload Manager: DB2 best practices: Implementing DB2 workload management in a data warehouse
What You’ll Learn:
Using a staged approach, this article guides you through the steps needed to implement the best practices workload management configuration on IBM DB2 for Linux, UNIX, and Windows with sufficient controls to help ensure a stable, predictable system for most data warehouse environments. This initial configuration is intended to be a good base for implementing additional tuning and configuration changes as needed, in order for you to achieve your specific workload management objectives.
You’ll be presented with a set of definitions representing the different stages of maturity for a workload management configuration in a DB2 for Linux, UNIX, and Windows database. These stages range from stage 0 through to the advanced stage 3 configuration. A specific configuration template and process is provided as part of these best practices to enable customers to progress from a stage 0 configuration to a stage 2 configuration. General descriptions and advice are also given about common stage 3 scenarios.
The article assumes a novice beginner and describes the individual steps and mechanisms at each point. A more experienced user can condense many of the listed steps to move from stage 1 to stage 2, making the transition in days of elapsed time rather than weeks as the suggested timeline indicates in a later section.
The steps outlined are focused on the efficiency of the system as a whole, regardless of where the work itself comes from. It is important to note that achieving the goal of a stable system might not necessarily also result in the achievement of any individual application service-level agreement (SLA) or specific performance objectives for queries. These more granular objectives might require subsequent changes to the workload management configuration, such as outlined in the section on stage 3 scenarios, which is outside the main scope of this document.
This is not a tutorial on DB2 workload management capabilities and does not attempt to provide comprehensive guidance in addressing all possible scenarios where DB2 workload management might be employed. It also does not cover all features within the DB2 product that might be of use in controlling resource consumption. The scope of this article is focused on describing the system stabilization approach in some detail and provides some general guidance for common advanced scenarios.
Paul Bird is a senior technical staff member (STSM) within the IBM Software Group development organization, sharing his time between the Optim and DB2 development organizations. Since 1991, he has worked on the inside of the DB2 for Linux, UNIX, and Windows product as a lead developer and architect with a focus on diverse areas such as workload management, monitoring, security, and general SQL processing. He recently became a member of the Optim development organization to expand his experiences. You can reach him at firstname.lastname@example.org.
Rimas Kalesnykas is a technical writer for DB2 for Linux, UNIX, and Windows. In the last 5 years, he was the documentation owner for a variety of DB2 subject areas, including the Command and API references, the Partitioning and Clustering Guide, Troubleshooting, and Workload Management. In 2008, he was a co-author of a best practices paper that describes how to improve data server utilization and management through virtualization. You can reach him at email@example.com.
Did you know?
See also my previous posts on Best Practices: