Since its inception, virtualization has changed the information technology landscape in many ways. With all the good that virtualization brings to the table, in some ways, it has made our jobs too efficient. One example is the ease and speed that we are able to deploy new servers. No longer are we waiting on physical hardware to arrive for a new deployment. We can “clone” our golden image in a matter of minutes and be on our way.
With the ease that virtualization brings to the table, it also introduces new issues into our day to day life as the administrator and care takers of the environment. Life Cycle Management can be one of those very issues. I have worked in several different types of environments over the years. One client, in particular, had life cycle management down to a science for any and all servers in the infrastructure. Every month, we would get the monthly decommission list of all virtual and physical servers slated for end of life. It took longer to fill out the paper work than to actually get rid of the server.In other environments, virtualization has helped life cycle management become more of a life cycle retention. One case in point was an old Windows NT 4 server that had an application running on it that could not be recovered. The people that had originally installed the application were long gone and so was the software needed to recover and/or reinstall. By the time I was made aware of this physical server, the hardware maintenance was expired and the server itself was showing signs of an imminent death. My task was to perform a physical to virtual migration before the physical server demise. Since this application could not be rebuilt, once the server was virtualized, I was to create several clones and backups of the virtual image to be used if the newly virtualized server ever crashed. There would be no excuse as to not being able to recover and restore this server now that it was virtualized. That was made very clear to me as I began this task.
Virtualization has made life too easy for us, or at least the management team, as they are now able to put off having to deal with replacing or upgrading a particular server in question and use virtualization to extend the life cycle of the application and servers as a rule going forward. As a matter of fact, once any physical server has been migrated to a virtual machine, the rush or need to focus money and/or resources on refreshing the applications sometimes seems to become more of an afterthought. After all, refreshing the physical hardware for a virtual machine is as easy as vMotioning the virtual machine to a new physical host.
VMware’s first attempt at addressing the life cycle management pain point was with a product called VMware Life Cycle Manager. Life Cycle Manager has evolved during its lifecycle and has now become a part of the VMware vCenter Orchestrator product. VMware vCenter Orchestrator is a development and process-automation platform that provides a library of extensible workflows to allow you to create and run automated, configurable processes to manage the VMware vCenter infrastructure as well as other VMware and third-party technologies.
The following list presents a quick look at the Orchestrator’s key features:
Persistence: Production grade external databases are used to store relevant information, such as processes, workflow states, and configuration information.
Central management: Orchestrator provides a central way to manage your processes. The application server-based platform, with full version history, allows you to have scripts and process-related primitives in one place. This way, you can avoid scripts without versioning and proper change control spread on your servers.
Check-pointing: Every step of a workflow is saved in the database, which allows you to restart the server without losing state and context. This feature is especially useful for long-running processes.
Versioning: All Orchestrator Platform objects have an associated version history. This feature allows basic change management when distributing processes to different project stages or locations.
One of the key things, I think, about the Orchestrator platform is its ability, and in some ways, expectation to tie in to 3rd-party products. Custom plug-ins can be applied and the continual use of the VMware API’s in the management of the virtual environment and the SOAP API’s to connect to the other 3rd party products. VMware has supplied a vast library of workflows to get most any environment started out of the box but also can be fully customizable to match the individual needs of any environment.
There are definitely other products out there that answer these needs but the point to make is that this need or pain point is sometimes fully overlooked or not implemented enough. Using lifecycle management, as an example, hopefully your environment is not the one that has ten year old legacy applications still around in the infrastructure.
Absolutely agree that lifecycle management is frequently overlooked in the virtual world; but it’s not just the application lifecycle as described in this blog. Because the life of VMs can range from minutes to years, if individual VM lifecycles are not effectively managed, the environment quickly becomes inefficient, with redundant VMs continuing to take up valuable resource and management time. Without effective lifecycle management, sprawl is inevitable.
Orchestrator is one way of dealing with it. But unfortunately not everyone has the time, expertise and interest in creating and managing individual scripts and workflows. For those who don’t, there are systems out there like Embotics V-Commander (www.embotics.com ) that provide policy based lifecycle management for the whole enterprise right out of the box.. with a few simple point and click setups and no scripts to manage. Even better systems like V-Commander simulate any policy you set up against the whole environment before it implements them… ensuring that the policies perform as you thought they would and not in some other manner. But no matter if you chose a do-it-yourself system like Orchestrator, or an out of the box system like V-Commander, it’s clear that lifecycle management is an essential requirement for any virtual environment.
Good point and tip David. No matter which method you choice it is important to have something in place.