By Dana Boehler
The speed of business has never been faster. Product release cycles have shrunk to timelines inconceivable in the past. Some fashion retailers are now releasing new product every two weeks, a cycle that historically only happened 4-8 times a year, and certain retailers even have product available immediately after it is displayed on the runway.
The demand for immediate insight into the state of sales numbers, ad campaigns, and other business functions has made the continuous aggregation of data commonplace. And if those factors weren’t pressure enough, the threat of ever-evolving security hazards is generating mountains of updates, code changes, and configuration adjustments — all of which need to be properly vetted before entering a production environment.
All of this activity needs to run on infrastructure that administrators like ourselves must manage, often with fewer coworkers to assist. Thankfully, for those of us running IBM i on IBM Power Systems, IBM has provided a robust cloud management tool that allows us to quickly spin up and spin down systems: PowerVC.
PowerVC allows users to manage existing IBM Power System partitions, create images from those partitions, and deploy new partitions based on those images. More recent versions of PowerVC support IBM i management and deployment (earlier versions did not).
Over the past year, I have been using PowerVC to greatly reduce the amount of time it takes to bring a system into the environment. Typically, creating a new system would take several hours of hands-on keyboard work over the course of a few days of hurry-up-and-wait time. The first time I deployed a partition from PowerVC, however, I was able to reduce that to about an hour, and after more refinements in my deployment, images, and process, I am now down to under 25 minutes. That’s 25 minutes to have a fully deployed, PTF’d system up and running.
The full implications of this may not be readily apparent. Obviously, net new systems can be deployed much more quickly. But more importantly, new modes of development can be more easily be supported. PowerVC supports self-service system provisioning, which enables teams to create their own systems for development, test, and QA purposes, and then tear them down when no longer needed. Since the systems are focused on the task at hand, they do not need the resources a fully utilized environment would need.
There’s more: Templates can be created in PowerVC to give the self-service users different CPU and memory configurations, and additional disk volumes can be requested as well. Post-provisioning scripts are supported for making configuration changes after a deployed system is created. In our environment, we are taking this a step further by integrating PowerVC with Red Hat’s Ansible automation software, which has given us greater flexibility in pre- and post-provisioning task automation.
In practice, using PowerVC removes many of the barriers to efficient development inherent in traditional system deployment models and permits continuous deployment strategies. Using PowerVC, a developer tasked with fixing a piece of code can spin up a clean test partition with the application and datasets already installed, create the new code fix, spin up a QA environment that has all the scripted tests available for testing the code, and then promote the code to production and delete the partitions that were used for development and testing.
You do have to make some changes to the environment in order to support this model. Code needs to be stored in a repository, so it can be kept in sync between all systems involved. The use of VIOS is also required. Additionally, note that when using this type of environment, the administrator’s role becomes more centered around image/snapshot maintenance (used for deployment templates) and automation scripting rather than the provisioning and maintenance of systems.
For full information on the product and its installation, I recommend visiting IBM’s knowledge center.