The company reached out to us at a difficult moment. The only IT person, who over the years had single handedly built and maintained the whole environment, handed in his notice, and with him practically all the knowledge of how the IT worked was leaving the company. There was no documentation, no network diagram and no list of what ran on which servers. On top of that, the physical servers that hosted the B2B store and the warehouse systems were nearing the end of their warranty.
The starting situation
The environment rested on the knowledge of a single person. Admin passwords were scattered, some access existed only in the memory of the departing employee, and users with problems called him directly, bypassing any process. The four physical servers differed in age and configuration, and two of them had warranties ending within a few months. A failure of any of them would mean the store going down, and nobody in the company could estimate how long restoring the service would take.
Why it was urgent
For a distributor operating in a B2B model, every hour the store and warehouse system are unavailable means real losses and irritated partners. The risk was twofold: ageing hardware out of warranty and the complete lack of any knowledge of how to restore the environment. In practice the company was one broken disk away from a serious problem, and the IT person’s departure left a gap that could not be filled overnight.
How we ran the takeover
We started with an inventory, still before the previous IT person left for good. We listed the servers, applications, access and the dependencies between them, recovered and organised the passwords in a password vault, and documented all of it. Only with a map of the environment could we plan a sensible modernisation, rather than working blind.
The next step was virtualisation on Proxmox. Instead of buying four new servers, we consolidated the environment into two hosts in a cluster. We moved the physical machines as virtual ones, so the company stopped depending on specific, ageing hardware. Snapshots gave us the ability to quickly roll back a change after an update, and restoring a machine from a copy is now a matter of minutes, not days of waiting for parts.
In parallel we launched a helpdesk with an SLA and a single point of contact. Instead of calling a specific person, users report problems to one place, and every ticket has a number, a priority and a history. We took over day to day support and the monitoring of hosts, copies and key services as part of ongoing outsourcing, so that the state of the environment is checked on a schedule, not only after a failure.
What changed
The company regained control of its own IT. The four physical servers now run as virtual machines on two hosts, which reduced rack usage and power draw by about 40 percent. Thanks to snapshots we restore a single service in a dozen or so minutes, not over days. The helpdesk responds to tickets within 30 minutes, so problems do not sit and pile up in a queue.
The most important change, however, is harder to measure. Knowledge of the environment no longer lives in the head of a single person but in current documentation. The departure of any administrator has stopped being a threat to business continuity, and the management knows who to call and what to expect when something goes wrong.