IT Hero Blog: Transitioning to a Hyper-converged Infrastructure: The Do’s and the Don’ts

For our IT Hero blog series, we’ve invited our customers to share their real-world experiences from deploying Windows Server software-defined solutions.

This month our IT Hero is Matt Lehmann. Matt is an information technology manager for Automotive Product Consultants (APC) Integrated Services Group, a leading marketing services company for the automotive industry. Matt has over 15 years’ experience managing information systems and as a network administrator.

Matt Lehmann
IT Manager – APC

Our company, APC ISG, provides marketing services to the automotive industry and operates a full-service mail facility and call center. We’ve developed a sophisticated data modeling program where vehicle data is analyzed and filtered to optimize results. It allows us to determine the correct timing and messaging for each customer. Simply put, our business model is dependent on having a data center that works. If it’s down, we lose money.

Storage Spaces Direct with Windows Server 2016 was a cost-effective solution for us and had the features we needed. However, when it comes to new technology, it’s never as simple or rosy as manufacturers or vendors spin it. Ultimately, us guys on the ground have to implement and manage the thing. Vendors are always going to tell you everything will work perfectly. But that’s usually not what actually happens.

I was lucky. DataON provided configuration scripts that made our HCI deployment easy, straight out of the box. They worked like magic. But you really need to know PowerShell, understand the scripts you are running and why they work or don’t. Eighty percent of the networking configurations in Storage Spaces Direct are done using PowerShell. Before deploying a live Storage Spaces Direct deployment, make sure to read, get hands on and go through the trial and error process. That was key for me.

If you get stuck, I suggest joining an active forum like the Storage Spaces Direct channel on Slack. Some of the guys on the forum are managing over a thousand active Storage Spaces clusters, and they’re extremely responsive. Not only can I get advice from experts, but I can see what problems other people are experiencing. It helps me better troubleshoot and learn from other people’s mistakes.

Always update Windows Server before deploying. I didn’t and there was a bug in that specific build that gave me all sorts of headaches. But before you do anything, make sure and check if there are any issues with the new update in forums.

I also used VM Fleet and benchmarked performance. Each time I made changes, I ran VM Fleet and checked the results against DataON’s base config. I simulated as many failures as I could and threw everything at the cluster it could or couldn’t handle. And then some. Afterward, I had a much better understanding of the do’s and don’ts, what tweaks worked, what didn’t, and I came away from it all the better.

Managing HCI is not the same as managing a SAN. But once you get over the learning curve, it’s worth it. After deploying our initial cluster, I feel comfortable with the technology and I’m adding a DataON Kepler two-node running Windows Server 2019 for a second cluster. I’m looking forward to 2019’s deduplication feature and some of the new performance increases.