Give in to data centre automation and change your life
Banish sysadmin drudgery
As an IT professional, unless you’ve been living under a rock you are probably familiar with automation, even if only in passing.
Automation process has been in use in the business world for many years, but somewhat paradoxically IT is usually the least automated department in any organisation.
Whole data centre automation and orchestration has been one of those topics that divide IT teams. Some speak of automation in the hushed tones reserved for discussing the holiest of holies.
Others find just speaking its name brings the sound of wailing and gnashing of teeth from the IT department.
Whole data centre orchestration has been pushed by vendors as the one true goal of any IT department in charge of web-scale infrastructure. In truth, orchestration is often poorly understood before, during and after any attempt at implementation. It needn’t be this way.
The traditional IT department model of the past 30-odd years has been silos. Systems manage the server infrastructure and try their level best not to talk to storage. Storage manage everything storage and often try to classify anything with more than 1GB of internal storage as part of their domain.
Networking administrators are seen as little more than semi-intelligent cable monkeys. More often than not the job is done these days by a pimply-faced youth who has yet to realise that systems is a better place to be.
I won't talk about the Exchange and unified comms teams. Those guys and gals are a breed apart.
Communication between these teams is often minimal at best, with each jealously guarding its own demarcated little fiefdom. In a modern world, as budgets and head counts shrink, responsibilities are being shouldered by staff already exceeding the 80 per cent threshold for a reasonable break-fix workload.
Whole data centre orchestration is about maximising operational predictability, security and maintainability through inter-team communication and IT process automation.
Teams can shift focus from a reactive maintenance cycle to working on projects that deliver business value
To put it simply, it is about automating every process involved in the day-to-day running of your data centre so teams can shift their focus from a reactive maintenance cycle to working on the projects that deliver greater business value.
If you are liking the sound of automation so far, then it is time to take a look at the tools available on the market to help us achieve these goals.
Knowing where to start when picking an automation solution (or combination of solutions) can be difficult unless you are familiar with DevOps.
El Reg has decided to make things easier by providing a brief overview of three of the major automation and orchestration software packages.
The first is probably the most popular of all the available offerings. Puppet has its roots in the open-source DevOps movement and has become a market leader in data centre orchestration and infrastructure management.
Whether you are managing just a few servers or a fleet of thousands, Puppet has the ability to scale. It offers a range of features that can automate every step of the software delivery process, including the provisioning of physical or virtual infrastructure, as well as test orchestration and reporting.
Focusing on reliability and stability, Puppet's features go beyond simple configuration management. Its greatest strength, though, is its training resources.
One of the biggest barriers to entry for those looking at automation and orchestration products is the often steep learning curve associated with creating runbooks. Puppet acknowledges this is a big issue and provides the training resources needed to tackle it.
No list of automation solutions would be complete without Microsoft’s Microsoft System Center 2012 Orchestrator. It is billed as a complete workflow management solution for the whole data centre, but it is with Windows Server infrastructure that Orchestrator really shines.
Chances are that if you are managing Windows Sever infrastructure above a certain scale you are familiar with other parts of the System Center suite. Like everything else, Orchestrator is reliant on other parts of the suite to deliver the most benefits.
Unless your budget is extremely constrained, this solution is well worth considering.
Chef’s offering differs fundamentally from Orchestrator and Puppet. Built with hybrid and hosted environments in mind, it has been embraced in a big way by those running wholly or partially hosted AWS clouds.
With an emphasis on cloud-centric features such as high availability, as well as zone-based replication and failover, Chef has put a lot of effort into catering for those running critical infrastructure in a cloud environment.
The company also offers privately hosted Chef for a premium subscription.
Follow the script
As the examples above show, whole data centre and hybrid cloud automation comes in a number of flavours and there are an entire ocean’s worth of resources available online that can further assist in deciding which tools are right for you.
Those tools are great, but what if you are not running a data centre or web-scale infrastructure? Can you still implement an automation and orchestration solution?
Are there some easy things you can do right now to show management that automation is not just for enterprises?
As an IT mercenary working with SMEs, it is not unusual for me to be supporting 15 to 20 servers and 300 to 500 endpoints spread across as many campuses. Like any IT professional supporting servers I handle many menial repetitive tasks which are easily orchestrated.
For example I hate it when end-users leave their workstations on over the weekend. Not only is it a waste of electricity (even with today’s reasonably green workstations), but it can lead to complications when you need to perform server maintenance.
Automating workstation shutdown via scripting is a simple yet effective answer and there are a number of ways it can be achieved.
- You can run a batch script to execute shutdown via command prompt as a local scheduled task.
- You can run a PowerShell script to execute shutdown as a local scheduled task.
- You can run a PowerShell script locally on a server and have it execute remotely on the endpoints of your choice.
- You can use IT process (runbook) automation or orchestration software.
Scripted automation provides a number of easy ways to perform simple tasks. Learning to automate administrative tasks on endpoints really is such small potatoes that junior help desk staff can take on some of the planning and implementation.
These small tasks are a great way both to learn about automation and to create examples that show management the benefits of automation in your clients’ workplaces.
Just the job
Whole data centre automation does a good job of covering the bases when it comes to web-scale, hybrid and fully hosted server infrastructures, but it is not always the best option when it comes to automating the everyday minor tasks that keep your endpoints humming.
While you can automate your servers with batch and PowerShell scripting, once your infrastructure grows beyond a certain level those tools won’t cut it. They don’t have the same power or flexibility to automate as Puppet, Chef or Microsoft System Center Orchestrator.
Choosing the balance point at which you switch between scripted automation and an automation package is a subjective decision. There is no specific moment where you absolutely have to move from one to the other. Generally, you will find out where that point is during the proof-of-concept stage of your project.
Automation is an integral tool for IT administrators as budgets and teams continue to shrink and infrastructures continue to grow in complexity, size and scope.
If you are running a hybrid or wholly hosted infrastructure, automation can save you an enormous amount of time and stress when it comes to monitoring and management. ®