How to use IBM Pureapplication System to achieve high availability during Operation maintenance

Brief introduction

Throughout the lifecycle of the application, operating systems, middleware, and applications need to be constantly updated, repaired, and patched. Pureapplication system provides an integrated maintenance management system that helps reduce the downtime of applications and the likelihood of human error, minimizing the time spent in repetitive maintenance operations. User interface and Automation tasks simplify administrative tasks that apply maintenance to multiple running systems.

This article is one of two articles on Pureapplication System multi-site implementation for high availability. Ideally, you should read another article using IBM pureapplication System to achieve high availability across multiple sites. This article describes the design principles of a high-availability stateless Web application, describes the design of a multi-site environment, and describes how application workloads deployed across these two sites provide high availability through a series of potential failure scenarios.

This article is written specifically for the architects and operations teams that manage Pureapplication System environment component updates. This article describes the methods provided in Pureapplication System to help maintain middleware and applications. These methods use automation and environmental knowledge to help minimize application downtime during the update process. This article provides a sample infrastructure that demonstrates options that can be used to further reduce the apparent disruption of application services. Finally, we will explore how application patterns, script packages, and images are updated and repaired to achieve synchronization across multiple systems.

Sample Infrastructure Overview

This article briefly describes the sample application infrastructure for deploying Pureapplication System across two sites, and provides high availability through a series of maintenance scenarios. As mentioned above, the design principles of a highly available stateless Web application are described in the first article. This article uses the same infrastructure to discuss maintenance technologies and equipment. For easy understanding, figure 1 is a hint for the infrastructure.

Figure 1. An overview of the sample infrastructure

In short, the Tradelite application will be deployed to a highly available IBM WebSphere application Server cluster running on both racks. These racks are located in geographically dispersed data centers and operate in "active-active" mode. We refer to the pureapplication system in "Data center A" as the primary partition, while the pureapplication system in "Data center B" is called a secondary partition so that it can be clearly identified in the discussion.

Now we continue to explain how to use the sample infrastructure to help minimize application downtime during maintenance. We will also discuss the types of updates that are frequently encountered in the life cycle of Web applications running on Pureapplication System. It then describes the process of applying these updates and helps maintain the high availability of your applications.

Maintenance of application Deployment

In applications deployed on Pureapplication System, you may encounter different types of updates for different components. The types of updates that are frequently encountered during the lifecycle maintenance operations of an application in the data center include:

Application Updates

Middleware Update

Schema, image, and script package updates

Other updates, such as operating system sharing and security patches

We will describe these update types in the context of pureapplication system and describe how pureapplication system helps to apply these updates seamlessly when leveraging the above infrastructure to help achieve high availability.

Application Updates

Application updates typically refer to a bug fix in an application or a new version of the application that needs to be deployed. In WEB applications, these updates are usually delivered in the form of a WAR file. Without disrupting the service, an administrator can use a device in the WebSphere cluster to push a wheeled upgrade (rolling upgrade) to the main frame. After the upgrade completes successfully, the same operation can be done on the WebSphere cluster running on the secondary rack.

Middleware Update

A middleware update is a fix to a middleware container for a managed application, or a new version of a middleware. In our scenario, the middleware container is the WebSphere application Server. We demonstrated how the integrated maintenance tool in Pureapplication System enables administrators to apply service updates to WebSphere nodes in the primary cluster, while user requests are handled by secondary racks. After the host has successfully obtained the service, the same process is performed on the secondary rack.

Two types of middleware updates must be considered: emergency fixes and service level updates. Emergency Repair is a temporary fix that handles minor problems with previous major versions. The service level update is the primary fix and the version upgrade of the new feature. In Pureapplication System, the processing of the two types of updates is different.

Updates to patterns, images, and script packages

The Pureapplication System leverages application deployment patterns. The patterns executed in pureapplication System make it possible to repeat the rapid deployment of cloud applications. This series of articles focuses on virtual system patterns, and you can find a further introduction to this pattern in the design of virtual system patterns.

You may need to apply updates to schema-related artifacts, such as images, scripts, and schema definitions themselves. The schema creator may make these updates and add new features to the schema, such as adding a new version of the application, if that version is part of the pattern, or fixing the problem with the current schema version, such as incorrect configuration parameters.

Synchronizing these artifacts is important in a multiple-rack environment running the same application, so the deployment on each rack is the same. We will help drive this synchronization process and promote the use of some of the tools that Pureapplication System provides to automate the task.