As an R&D lab manager, I love talking about automation with customers. Not only do they always provide us with amazing insights about what they need in solutions, but I always leave the conversation with something new to think about.
Last week, for example, we were meeting a client organization that is just starting to automate aspects of their IT environments, and we were discussing how they will be able to use automation to do much of what they do today, including manage databases mounted on virtual machines. Suddenly, their architect asks, "How does DMA handle Golden Images?"
Of course, hearing that term took me back a few years, but it also got me thinking about how Images and Automation can work together to achieve what I think of as a Golden Recipe of better lifecycle management of databases, particularly those on virtual machines.
The limits of image-based lifecycle management
As you well know, it’s pretty straightforward to snapshot a virtual image of, say, an Oracle database on a single system, and then make paths with vCloud Director to spin up those images anytime someone orders the database. And that works pretty well for a simple dev cloud. But in production environments, things get more complicated, especially when multiple application teams start making modifications.
You can only take image management so far. In just about any scenario that’s more complex than an Oracle database on a single system, managing lifecycles with images requires the kind of time and effort that can be better achieved through automation. The fact is, you probably need a blended approach of both images and automation to manage lifecycles, and this is where the Golden Recipe comes in.
Cooking up the perfect batch
The Golden Recipe model is pretty simple: A repeatable deployment of applications requires the use of both images and automation. As IT organizations build Hybrid Clouds, one of the early questions they face is, What is a “best practice” value for X and Y? Experience varies, so let us know your thoughts on this equation by posting a comment at the end of this blog.
Where image-based approaches and automation meet
In my experience, one clear sweet spot for an image-based approach is with components that are foundational across application team needs. For example, it is not uncommon for an enterprise to have standardized on specific RHEL images, possibly coupled with component provisioning for database or middleware applications.
But image management introduces complexity when images are incorporated into a broader service design. For example, complex HA/DR PaaS configurations such as Oracle RAC require multiple IaaS provisioned systems to be weaved together. In these cases, automation can provide the right level of flexibility to disconnect the Oracle RAC configuration from the images themselves.
Therefore as part of a broader architecture, practitioners end up simplifying the images to the lowest common denominator across application teams and then use automation services to create the clustering configurations like Oracle RAC
Automating the last mile of configuration
A big part of how automation adds value to the Golden Recipe is in its ability to capture application-specific requirements of dynamic infrastructure. We sometimes refer to this as the last mile configuration, and it comes with additional complexity in the application-specific information that should flow with the application as it is promoted from dev to test to production. The dynamic nature of production environments lends itself more to capturing the application specific configuration in targeted content like DMA instead of image reconfiguration, which could lead to dramatic increase in the number of images that must be managed.
Tracking the compliance state
In production environments, hybrid data centers require compliance scanning and remediation of application infrastructure. Automation can enable organizations to drive remediation actions via compliance reports and achieve new levels of visibility into database and application server estates. If you are already provisioning databases with DMA, for example, you can use the solution to visualize DB/AppServer estates and take actions to report and remediate in a blended environment.
I would like to hear how your organization balances image-based approaches to managing the database lifecycle with automation—what do you think the Golden Recipe is? Post a comment below and let us know.
HP DMA can help with this lifecycle management of tasks across databases mounted on physical servers and virtual machines. To find out more: