Say you run an automotive supplier in Novo mesto with seventy employees. Five machines of different ages hum in the server room: the ERP with its database, a server collecting data from production equipment, a terminal server for home working, a file server with drawings and control plans, and an old PC wired through a serial port to a measuring instrument in quality control. Your customer insists that inspection data stays available for years, and any server outage means a stopped line or a late shipment. Most of the machines are half idle, yet each has its own power supply, its own disks and its own warranty, two of which have already lapsed. Virtualisation consolidates these servers onto two or three hosts. If a host fails, its virtual machines restart on another, and a snapshot is taken before every update. Because planning matters more than installation, a large share of the work happens before you buy anything. It is all done remotely: you order the hardware from your own supplier, the supplier or a local contractor racks it, and we install, migrate and test.
With virtualisation, most of the savings come from sound decisions before the order is placed. These are the questions we answer before you sign anything.
Instruments on serial ports, USB licence dongles and software tied to specific hardware are frequent exceptions. We ask the vendor whether pass-through is supported; if not, that PC stays physical and we simply protect it.
For a week we record CPU, memory, disk and network load, month-end included. Host size follows from that, with headroom for growth and for the day one host is out of action.
Hyper-V when the estate is all Windows, Proxmox VE when cost and openness come first, VMware when the software vendor requires it. You receive a written comparison of the trade-offs in your situation.
Standard edition, with every host core licensed, covers two Windows virtual servers; Datacenter covers any number. We work out which pays off for your VM count and give a three-year cost including the platform and backup software.
Replicated local SSDs, shared storage, or a Ceph cluster with three hosts. Management, replication and user traffic each get their own network.
Copies go to a separate device and to EU cloud storage, never onto the same host. At least one copy is immutable, so ransomware cannot encrypt it.
Servers move one at a time in agreed windows. At the end we deliberately shut down a host and time how quickly the machines come back elsewhere.
How long it takes depends mainly on the number of servers and the hardware delivery time. No site visit is needed.
Inventory, a week of measurements, a map of links between systems and a talk about how long each one can be down.
Platform, a hardware list for your supplier, network plan, backup design and the three-year cost.
Once a local contractor has racked the hardware, we install the platform, configure networking and storage and check performance.
Servers move in sequence, the failover drill follows, and you receive the recovery plan and documentation.
A snapshot is not a backup. Snapshots are handy before an update because they let you roll back in minutes. But they live on the same host and the same storage as the virtual machine. If the disk array fails or ransomware encrypts the datastore, machines and snapshots vanish together. We use snapshots only short term; real protection comes from copies on a separate device and in the cloud, restored on a regular schedule.
You can, as long as you accept that if that host fails, every server stops. Where the budget will not stretch to two hosts, we set up fast-restoring backups, a pre-agreed quote for replacement hardware and a plan to run key machines in the cloud temporarily. The accepted risk is put in writing.
Usually a few hours per server, outside working hours or at the weekend. In shift-based production we align the moves with the production schedule and with any annual shutdown. You get the outage timetable in writing beforehand.
Yes, provided someone who knows it is running it. Proxmox VE is open source, and its developer offers paid subscriptions with stable updates and support, which we recommend for production use.
Not necessarily right away. Given the changes to VMware licensing, we compare the renewal cost with Hyper-V and Proxmox and check whether your backup tool and specialist software support another platform. If you do switch, it can happen in stages.
Often only part of it. Virtual machines can run with providers such as Hetzner, OVHcloud and IONOS or in the EU regions of Azure and AWS. Systems wired to production equipment or measuring instruments usually stay on site. We cost both routes side by side.
Which servers are running today, what do they host, and are any special devices attached? We will answer with initial questions and a rough design.
Enquiry received
An engineer will write back no later than the following working day. Faults that leave your staff unable to work are always looked at first.
No town by that name here. Check the letters, or settle for the nearest larger place. Our engineers work only over the internet, so Kočevje and Koper get precisely the same treatment.
