Optimizing IBMDB2pureScale transaction throughput in virtualization

Nowadays, hardware virtualization is prevalent in the entire IT industry. Using one server to execute the workload that can be completed by several servers is an excellent value orientation for enterprises of any scale. Although there are a large number of documents (online and printed documents) available, it is difficult to control the vast amount of resources.

Nowadays, hardware virtualization is prevalent in the entire IT industry. Using one server to execute the workload that can be completed by several servers is an excellent value orientation for enterprises of any scale. Although there are a large number of documents (online and printed documents) available, it is difficult to control the vast amount of resources.

This article will show you how to address this challenge by providing a series of advanced, planning, and deployment best practices, as well as a series of simple optimization steps, using a sample environment for IBM POWER7®Run an IBM DB2®PureScale®.

Today, database servers may be implemented on a virtual machine (VM), where processors, memory, disks, and network components are allocated from a shared pool of available resources on the physical machine. This allows you to integrate multiple workloads into a single server. Although large physical machines may require higher upfront costs, integrating many workloads into one server means more efficient use of system resources, this helps reduce power consumption, reduce data center space occupation, and ultimately achieve a lower total cost of ownership (TOC ).

However, this paradigm introduces a coordination mechanism (also the essence of this paradigm) in physical machines ), you can allocate, map, and manage physical resources of virtual hosts or logical partitions (LPAR) on IBM Power Systems. The implementation of this mechanism (usually called Virtual Machine Hypervisor) provides a robust, scalable, and manageable virtualization solution that helps minimize the ROI of customers ).

The hypervisor is IBM PowerVM.™As part of the virtualization solution, you can obtain the LPAR configuration files and deploy them to IBM Power Systems in an efficient and seamless manner. It can virtualize multiple operating systems (OS) and has scalability. it can deploy up to thousands of VMS on one server, A complete set of tools can also be integrated to quickly and conveniently plan, deploy, and manage LPAR.

For example, using the IBM PowerVM solution, you can not only set the number of resources received by the LPAR, but also control these resources (such as processors and memory) (from the same interface) sharing methods between LPAR of other systems. If there are some less important LPAR, they are often idling, it is best to implement a shared processing LPAR configuration file. This means that unused processing resources can be placed in the shared processing pool for use by other LPAR. If the LPAR is mission-critical, although it is sometimes idle, you need to set a dedicated processing configuration file to process resources strictly coupled with the LPAR, and other LPARs are unavailable.

There are a large number of virtualization literature available (see the following references), from short articles to comprehensive books; locking in various scenarios, from specific database and application middleware workloads to general host deployment policies. This article will provide a brief list of LPAR best practices for planning and deploying, and then use a sample database environment in IBM POWER7 through a series of optimization steps®Run an IBM DB2®PureScale®The installation demonstrates the advantages of such best practices.

Best practices

To allow physical servers to provide additional resources for additional VMS and optimize hardware usage to maximize server performance, the demand is growing. The following best practices and suggestions can help you ensure the implementation and maintenance of optimal configurations.

1. understand and record server intentions, purposes, and objectives

Record server goals not only help system architects and administrators follow their plans on a limited list, but also provide a document for new teams or members to refer to when using the server in the future. Even if the usage mode is not constant, maintaining a dynamic document using server protocols by the maintenance team is also very valuable. It is easy to use the IBM System Planning Tool (SPT) to complete this step.

2. ensure that the system firmware is up to date

To obtain the best-performing hypervisor, install a system with the latest firmware version for use. If you have set a Hardware Management Console (HMC), you can use its maintenance function to update it.

Figure 1. after selecting the server, open the "Updates" task to view the "Upgrade Licensed Internal Code to a new release" option.

If HMC cannot be used to update servers, another method is available. with IBM support for manual patching, the site also provides patches and updates for other components (such as HMC or SPT.