I talk to a lot of IT leaders in manufacturing who are running critical production workloads on hardware they know is living on borrowed time. The VMware hosts are old, the SAN is past end of life, and the support contracts have lapsed or are about to. The plan, such as it is, amounts to keeping it running until it stops.
I understand why. Manufacturing environments rarely get a maintenance window every weekend, and production keeps running regardless of what IT wants to modernize. So the infrastructure conversation keeps getting pushed to next quarter, then next year, until something fails and the conversation happens anyway, under much worse conditions.
We recently worked with a manufacturing customer who decided to move before that moment arrived.
The Starting Point
Their environment was familiar to anyone who has managed infrastructure in a plant for a decade or more: a handful of aging VMware hosts running the full spread of production and business systems, backed by a SAN that had reached end of life, with vendor support and firmware updates both ended. Every day it stayed online was a day past its intended life.
There was no single dramatic failure driving the project. It was the accumulation of risk. Leadership teams eventually price out what an unplanned outage costs on a manufacturing floor, and the figure is large.
What We Did
Our professional services team led the full engagement, from design through migration through go-live.
On the compute side, we moved them off VMware and onto a Proxmox VE cluster, three nodes running on Dell rack servers. That decision gets a lot of attention right now given where VMware licensing has gone since the Broadcom acquisition, and licensing was one of several reasons. Proxmox VE gave this customer a modern, well-supported hypervisor platform without the multi-year lock-in and unpredictable renewal pricing that so many of our customers are trying to get out from under.
NetApp handled the storage side, for reasons that come down to the size of the IT team.
The NetApp Piece
We designed and deployed a NetApp AFF A20 as the new shared storage backbone, running ONTAP One and connected over 25GbE, serving the Proxmox VE cluster over NFS. Five years of SupportEdge coverage came with it, so this customer avoids a repeat of the end of life problem in year four.
What NetApp brings to an engagement like this goes beyond capacity: the operational tooling around the storage, the snapshot and replication capabilities, and the efficiency features that turn raw terabytes into a platform that makes backup, disaster recovery and day to day operations manageable for a lean IT team. For a manufacturer without a bench of dedicated storage engineers, the operating tools count for as much as the hardware spec sheet.
The migration itself is where the storage choice paid off first. We used the NetApp Shift Toolkit to convert the VMware virtual machines onto the Proxmox VE cluster, and because each conversion ran against a FlexClone copy on the array, it finished in minutes instead of hours. There was no long copy window per workload and no second set of storage to stage everything on. For a plant with no spare weekend to give us, shrinking the conversion time per machine is what made the schedule realistic, and it took a sizable line item out of the cost of the move.
The same array carries defenses the old SAN had no way to offer. Snapshots run per volume on a schedule and are retained on the array, so a bad change or an encrypted share has a recovery point measured in minutes. We also configured on-box ransomware detection: ONTAP watches for the workload behavior that signals an encryption event, raises an alert, and takes an automatic snapshot the moment it sees one. That response happens on the storage, underneath the hypervisor, so it keeps working even when something above it has already been compromised.
This is the kind of pairing we like recommending: a modern, cost-predictable hypervisor layer on Proxmox VE, sitting on top of enterprise grade storage from NetApp that runs without a dedicated specialist.
What the Savings Paid For
Here is the part of this engagement I like most. Before the refresh, this customer had no disaster recovery at all, and that was a budget outcome rather than an oversight. Renewals and support costs consumed the infrastructure budget every year, and DR was the line item that kept losing.
Moving off per-core VMware licensing changed what they could afford. The savings covered the refresh and left room for the capability they had wanted all along, so we replicated the new Proxmox VE cluster to ModernOps DR as a Service. They now hold an offsite copy in our data center, a documented recovery procedure, and scheduled restore tests against it.
The same budget that was disappearing into renewals now pays for a modern platform, on-box ransomware defense, and a recovery site. They came out of this with more capability than they went in with, at a cost they can predict. The wider version of that trade, where the gear leaving production becomes the DR site itself, is the second-job model.
We See This Pattern at Other Manufacturers
This is a pattern we have executed for a number of our manufacturing and industrial customers over the past year. Aging VMware environments and end of life SANs are among the most common risk profiles we see when we walk into a new account, and the technology fix is rarely complicated. Deciding to act before the hardware decides for you takes longer.
If your infrastructure conversation has been “it’s still running” for longer than you would like to admit, that is usually the signal to have a different conversation. We have done this enough times to know what a good outcome looks like, and we would rather help you plan it than help you recover from it.
If you are running infrastructure like this, talk to us before it becomes an emergency.