32 просмотра
Оглавление
How I Built a Private Cloud: Experience and Architecture
Disclaimer
This publication may contain strong language and materials that may be inappropriate for some readers. The content is intended for readers aged 27 and older. If you are sensitive to harsh expressions or similar elements, we recommend refraining from reading further.
The author and editorial team are not responsible for any negative reactions caused by the use of profanity in this text. The use of such language serves artistic or stylistic purposes and is not intended to offend or humiliate anyone.
You read this at your own risk. All characters and events are real. All stunts performed by professionals.
Introduction
At the end of 2021, I was offered a project to build a private cloud. The key question was simple: “Do you know Ruby on Rails?” — I did. I agreed without much hesitation.

Phase One: At First, It Was Interesting
It all started with meeting a lazy but experienced sysadmin, an architect, and a senior developer. They seemed kind and fun, and we began exploring an open-source concoction from Red Hat that we had chosen as the foundation for our cloud. We examined the code, the architecture, the provided frameworks, and proposed solutions. I suggested building a client-facing control panel, which was accepted without objections.
On my end, I assembled an MVP of the portal, the integrator wrote the part responsible for vmWare integration, and after a successful demo, both our leadership and the integrator’s team high-fived. But to me, it was clear: this was only the beginning. After hiring a frontend dev, we continued developing the cloud and integrating the system into production.
Import Substitution and Stress Testing
Then came the sanctions, and the next stage involved testing the "import-substituted" virtualization management system. Hiring was happening in parallel, but since I wasn’t involved in it, unfit candidates were brought in — and left just as quickly.
By summer 2023, new leadership pushed through an inappropriate hire. Around this time, our frontend dev left, and the team started being “built” from random people.
Then — A Zoo
Juniors with unexpected habits began to fill the team. One couldn’t hear tasks, another barely spoke at all. The sysadmin burned out and quit, while on paper, needs were being “met” and an illusion of vigorous activity was created.
Meanwhile, the infrastructure was growing. But team communication had completely broken down.
Autumn: Contract Ended, No Result
By the end of the contractor’s term, less than half of the deliverables from the original spec were implemented. There was no internal specification at all. Time was lost, the project stagnated, and the team was demotivated. A copy of “7 Habits of Highly Effective People” collected dust on a shelf. Expectation vs. reality.
Debriefing
After the New Year break, attempts to stir conflict became a management tactic — a mass one. Work ground to a halt, and trust hit zero. That shaped the policy of future interactions.
Straight to Hell and Back
When I landed, I realized: this was no longer a technical project — it had become a theater of traumatized managers. Once ignored and ridiculed, they now clung to power, acting out through control, fear, and sabotage. They understood neither people nor systems, but relished the right to sign things, as if that made them competent. Everything they touched turned to dust — as happens in any structure where hierarchy is valued over results.
These are the first managers I’ve met who don’t even spark the desire to argue or prove anything. You just want to remove them from the system before they finish destroying it. Their path is that of wounded doers given power without knowing how to use it. They can’t build or lead — but they excel at missing deadlines, throwing colleagues under the bus, and trampling anything that doesn’t fit their narrow worldview.
This isn’t management. This is Darwinism with an Agile handbook for the clueless. The survivors aren’t the strongest — just the most flexible in chaos, the most shameless in faking work, and the quickest to shift blame. These people turn processes into manipulation arenas, fear into a tool, and their incompetence into a norm.
When such madness takes over a project, there aren’t many options left: document what’s happening, watch your back, and don’t let yourself get pulled under. Then — stand up, brush off, and start building something of your own. Without clowns and ornamental “leaders.” Without motivational lies. With an engineering approach that has a spine, not just a slideshow.
A system where people aren’t afraid to think, and decisions are born not in the trenches of toxic politics, but in dialogue — among those who actually know what they’re doing.
On the Edge
From the outside, it might seem the system is working — over 5,000 virtual machines were created in 3.5 years, with almost no critical outages. But in reality, it’s an illusion of stability. There’s no monitoring, no metrics, no architecture. Everything runs on grit and memory. Any attempt at improvement is drowned in bureaucracy or dismissed by those who can’t tell the difference between a button and infrastructure. They tried to “replace” me with a PHP dev — apparently thinking clouds grow out of index.php
. Bugs are fixed as isolated cases, not as systemic errors. Engineering thinking has been replaced by office blindness. No one wants to repair — slow decay is easier.
The team assembled to support and develop the Private Cloud, unfortunately, was formed without a shared engineering strategy or a clear understanding of roles and responsibilities. As a result, some specialists remain out of inertia, others lack the competencies for the project’s maturity level. Inter-team collaboration is minimal, regular meetings are nonexistent, and tasks are broken down without architectural or functional context. This isn’t the fault of individual contributors — it’s the result of absent system-level leadership and engineering culture. What’s being maintained is more in spite of the system than because of it.
The King’s Move
The world holds its breath.
China extends a hand, as if offering to sign a new universal religion — an agreement on AI rules. This isn’t just technical: it’s a new narrative, a myth of the future trying to lock itself into place.
In our case, the king turned out to be a clown, listening to priests of his own invention. I proposed my own faith — engineering. A system of verifiability, load, fault tolerance, and causality.