It’s been a year and a half since the last article in this series. Now it’s time to move on to phase two and bring the project closer to its full realization.
Although the material for this second phase is rather sparse, it holds exceptional importance for the project as a whole and plays a pivotal role in its implementation. Moreover, this article can also serve as a kind of “mission debrief,” summarizing the results of the server room’s operation over a significant period.
The general takeaway so far is that the decision to set up the server room in this specific location and manner was an excellent one. The best proof of this is the fact that, throughout this entire time, it has required absolutely no attention. None at all! The server room and the equipment housed within have operated entirely autonomously, without needing the slightest intervention in their work:
Just as no one ever paid much attention to the unused space above the kitchen cabinets, over time, everyone forgot about the small cabinet with the glass door installed there. This proves that the choice of location for the server room was spot on — it doesn’t bother anyone or anything, asks for nothing in return, yet provides immense utility. Few of my projects can boast the same level of unobtrusiveness and effectiveness.
At first, I was a bit annoyed by the high placement near the ceiling, as it meant that accessing the server cabinet required more than just a chair — I had to lug in a full-sized ladder. And it’s heavy. However, in hindsight, this downside turned out to be negligible. So far, I’ve only had to climb up there once! That was when the NAS flagged an issue with one of the RAID (RAID5/RAIDZ1) drives. I had to replace the disk, but fortunately, I always keep a spare, unopened drive of the required model on hand for such emergencies.
Interestingly enough, the “problematic” drive turned out to be mostly fine. SMART diagnostics did report a dozen or so slow read/write cycles, but no new issues have arisen since. The disk passes all tests, including stress tests. I have no idea what the system thought it detected. For now, I’ve set it aside with a label that essentially says, “Hmm…” — though I can guarantee it will never make its way back into the server.
When it comes to server reliability, there’s no room for risks. Whether the system imagined an issue or not, my server holds literally my entire life: family photos dating back to the 1970s, all the materials from every project (including backups of blogs spanning years), favorite movies and shows (like Cobra, Commando, Police Academy, Total Recall, and so on), a software and game archive (from arj.exe and DOS Navigator to Fallout 4 saves I still haven’t finished), and, well, all kinds of other treasures — from horse-related curiosities to cat memes and receipts from purchases. Sure, I have independent backups of all this, but I take every single NAS notification very seriously.
“There are two types of people —
those who make backups
and those who now make backups“
— The modern wisdom.
The only disappointment was that not all elements of the home infrastructure were centralized in the server room. The internet provider’s modem and the primary router still resided in my study, which is in a completely different part of the house.
Relocating them to the server room was hindered by a peculiar choice made long ago — before my time — regarding how the internet cable was routed into the house:
The “In” point is located on my neighbor’s property — he serves as the Keeper of the Switch Box for our group of houses.
So, here’s the setup… The cable runs from this “box” along the entire far perimeter of my lot (covering an area significantly larger than the one roughly marked on the model above), then climbs the exterior wall of the garage to the attic, and finally drops down to the “Out” point in my study.
This arrangement made sense back in the late ’90s when the internet was delivered over a phone line. All the local home telephony infrastructure was bundled together with the electrical wiring on a single side of the house. That part of the telephone infrastructure is still there today (a gray box mounted on the pipe below the electric meter):
I assume that at the very beginning, when the subdivision was being developed, a deep trench was dug around the perimeter of all the involved lots using a tractor, so that all the wiring could be conveniently routed to multiple homes. At that stage, everything was fine — reliable, logical, and adequate.
However, years later, as the internet became a household necessity, things started going off the rails. Naturally, no one brought a tractor back to dig a deep trench or lay new cables in an organized manner for every house. Instead, it became a matter of private, ad hoc solutions for each individual home. The new cable was often laid just below the surface, covered with only a couple of inches of soil. And if it was particularly hot outside, the cable might not even have been buried — just stomped lightly into the ground. It’s safe to say that no one was trying to reinvent the wheel here. They simply followed the principle of “if the phone cable went this way before, I’ll bury the new cable roughly the same way.”
Worse still, the installer might have thought, “Why bother going around the corners? Why not just cut across the lawn?” A shortcut like that would practically beg to be severed by a shovel during spring gardening — and indeed, that happened more than once. Each time, a new technician would come, have the same idea, and repeat the process, leaving the new cable only lightly dusted with soil.
Based on my archaeological digs in the lawn, this happened at least three times — I’ve uncovered three separate coaxial cables, each hacked to varying degrees by shovels over the years. Fifteen years ago, when I was setting up internet for myself, I personally oversaw the process to ensure the installer didn’t get any “creative” ideas. But at that time, I wasn’t yet thinking in terms of organizing a server room in the house. So, while the cable no longer cut across the lawn, it still followed the old phone line route all the way around the property to the garage.
That was the initial state of affairs in this part of the project. Now, here’s how I decided to break this vicious cycle once and for all.
First, taking into account the new requirements and the current location of internet access points in the house, I decided to lay the cable along a much shorter and more efficient route directly to the server room:
Yes, as before, the cable would have to come out of the ground and run up the exterior wall of the house to reach the attic. But since the server room is practically in the attic itself, that’s now the most logical endpoint for it.
Second, the cable needed to be buried at a depth sufficient to protect it from an unsuspecting shovel. Not the full 4 feet into the ground, as required by local regulations for residential electrical wiring, but at least 1.5 to 2 feet. At that depth, it’s much less likely to be accidentally hit during routine yard work.
Third, I needed to ensure that the cable could be replaced in the future without requiring excavation. Here’s the issue: despite my repeated pleas to various providers, they stubbornly refuse to bring fiber optics to our subdivision. Their reasoning? “Not economically feasible.” In our sleepy cul-de-sac of just a few dozen homes, it seems I’m the only one actively asking for it. And providers aren’t willing to deploy their tractors and cable-laying equipment just for me.
But sooner or later, I’ll make it happen. I’ve been systematically spreading the gospel of fast internet among my senior neighbors, regularly submitting requests, pestering providers, writing to the HOA, and so on. One day, it’ll happen. And when it does, the coaxial cable will need to be replaced with fiber. I don’t want it to be buried haphazardly near the surface again.
So, the cable must be laid underground in a dedicated conduit. This way, when the time comes, it’ll be a simple matter of “pull one out, push another in” — no digging required.
Re-routing the internet cable according to these conditions became the primary task for the second phase of this project.
At this point, two projects converged into one: the lawn irrigation system restoration and the internet cable re-routing. Both required trenches to be dug, one for the sprinklers and the other for the cable. While the irrigation project will have its own dedicated article, this one will cover the points they share in the context of the server room.
The digging had to be done from the fence to the house. If you take a look from a bird’s-eye view using Google Maps, it would look something like this:
From a human perspective, this trench runs along the entire section of the fence, starting from the stump visible on the right side of the frame and extending just as far beyond the left edge of the image:
Ignore the fallen fence section and scattered plants caused by a recent hurricane — everything has long since been repaired and restored.
The trench begins here, in the corner of our lot:
Right where the trench meets the fence is the location of the shared communications box on the neighbor’s property.
At this stage, everything was smooth sailing. The trench was dug to the desired depth, and pipes were laid along the way. These included water pipes for the irrigation project and “electrical” conduits for this internet cable project:
This excavation extravaganza continued without a hitch — that is, until the trench reached the house. At the same time, it also reached the enormous, towering pine tree growing next to the house. The last survivor of its kind.
Originally, we had five pines and a massive oak in the backyard. Over the past fifteen years, four of the pines fell victim to lightning strikes at different times, and the oak was uprooted by a hurricane about five years ago. We still miss those trees — they were beautiful and created the illusion of living in a forest. That made it all the more important for us to preserve the last remaining pine.
Granted, it’s not exactly a picture-perfect example of paradise, but it’s ours (the second pine in the background is on the neighbor’s property):
Hurricane Beryl, which snapped the fence in the photo above, also managed to break off a massive branch at the very top of this pine tree:
Incidentally, that broken branch is still hanging there as of December 2024. We’re just waiting for it to either fall on our heads or smash through the roof. Just another day in the subtropics — nothing out of the ordinary…
Anyway, during the excavation process, the pine tree did this to the trench:
Cutting through all these roots was out of the question. The trench was too close to the tree, and such an operation could kill it. Even if it survived, it would lose a significant portion of its “anchors” that hold it to the earth. And come next year’s hurricane, we really don’t want to be worrying about whether we’ll be crushed by a large, dried-out branch or an entire tree, for heaven’s sake!
So, in this section, the trench wasn’t dug with a shovel but rather excavated using a “hydro-abrasive” method with an ancient power washer… or a “Kärcher”… or whatever… You know, one of those contraptions that sprays water at high pressure — the kind you use to clean concrete pathways, wash years of grime off cars, and all that good stuff:
Using the pressure washer, I worked along the trench route, searching for gaps between the roots. The soil was washed out from these gaps and then scooped out from the open section of the trench. The idea was to create a pathway where I could slide the conduit with the cable underneath all this chaos.
Of course, this couldn’t go on forever. The pine is massive (see the photo above). Its roots extend not only within the trench and along the wall of the house but far beyond my property. And the deeper you go, the thicker the roots become, with less and less space between them. The capabilities of the pressure washer in this scenario were far from unlimited. It’s not exactly designed for this kind of task — to put it mildly.
After almost a week of struggling in this area, I gave up and decided to redirect the trench toward the house wall, opting to run the “electrical” conduit above ground along the house instead:
This entire section proved impossible to dig or wash through using the tools (and strength) at my disposal:
So now, the conduit with the cable runs along the house wall and then turns up into the attic. At the point where the conduit transitions from underground to above ground, I had to install an electrical junction box (visible as the round fixture in the bottom-left corner of the photo). This was necessary because it’s impossible to push a cable through such a sharp transition in one go. The cable needs to be caught and pushed further separately into the segment leading to the attic:
In the end, it didn’t turn out as perfectly as I had envisioned at the start, but no DIY project is complete without some adventures. It could have been worse… I don’t know… maybe an oil geyser could have erupted from the dig. Or man-eating ants might have attacked from the depths of the lawn. Two feet deep is no joke! You could even trip and sprain an ankle. That’s no fun either.
The final challenge was connecting the new cable to the switch box. The problem here is that I’m not legally allowed to mess with the communications box without a license, even though it doesn’t even have a lock. It’s just a couple of unassuming covers placed over simple frames with connectors sticking out of small ground-level wells:
But such is the law… Accordingly, I couldn’t connect my cable directly to the contacts without risking the wrath of the gods. You can’t see it in the photo (I took it by leaning over the fence), but the sides of those covers are inscribed with ominous runes warning of all the heavenly punishments awaiting anyone bold enough to tamper with them.
Honestly, calling and waiting a week for certified technicians to show up while sitting next to a trench in the middle of the lawn was more than I could bear at that point. Thankfully, a “repair kit” — consisting of a coaxial cable stripper, crimper, couplers, and connectors — is considered a must-have tool in any homeowner’s arsenal around here:
So, I simply took the original cable from the ground (there it is, the orange one) and spliced it to my new one (the black cable from the conduit) using a coupler:
Cutting your own cable with a shovel on your lawn isn’t against the law. Nor is repairing it yourself. In fact, around here, this is a totally routine occurrence, happening in practically every other yard.
And so, it happened inside yet another electrical junction box:
Yes, I cut corners! Yes, I should have called the technicians and routed the cable directly to the box to avoid adding a break in the line that could introduce interference. But I was desperate to finish these excavations as quickly as possible. I was so unbelievably fed up with digging through trenches and wrestling with roots that a single coupler on the cable didn’t seem like such a mortal sin. All I wanted was to bury everything back and forget about it like a bad dream.
In my defense, the original old cable running along all the fences had, over the years, been dug up and chewed through by the Dog multiple times. The number of couplers on that line was practically train-length. And yet, none of the breaks or interference ever stopped me from getting my promised “Download speeds: up to 500 Mbps; Upload speeds: up to 100 Mbps” from the provider. Well, give or take a little (it always reliably hit at least 400 Mbps). So, reducing the number of breaks in the line to just one clearly isn’t something to lose sleep over.
Finally, all the trenches were filled back in, and I could start forgetting this whole nightmare with shovels:
By spring, the grass will have covered up all the remaining mess, and this dreadful memory will live on only in the photos archived on the server.
As the final act, the modem and router were ceremoniously removed from my study and installed in their rightful place in the server room:
The current setup of the server room looks like this:
- Modem: Since the modem technically belongs to the provider and tends to be replaced whenever they upgrade their services, I don’t bother with custom mounts. Each new modem comes in a completely different shape and size (not standardized server equipment), so it quietly sits in a corner, ready for its next replacement.
- Router + Wireless Access Point: This has been serving me well for years, so it’s mounted on a swivel bracket. This setup allows the antennas to face outward while making it easy to tilt the router out of the way when access to the equipment behind it is needed (as seen tilted to the side in the right-hand photo above).
- UPS (Uninterruptible Power Supply): Keeps everything in the server room running during a power outage for about 30–60 minutes, depending on the load. On average, I can count on around 40 minutes of backup power.
- NAS (TrueNAS): The workhorse of the server room. No shaking, no stirring.
- Central Power Supply for Security Cameras: All the cameras around the house are wired to this unit. Since everything in the server room is connected to the UPS, the cameras continue recording for a while even during a complete power outage.
- Network Hub: Only a few devices are connected wirelessly to the router, such as my wife’s Chromebook, our phones, and countless smart home controllers. Everything else (TVs, security cameras, my main computer, my daughter’s computer, the printer, home theater, etc.) is hardwired through this switch. Some rooms have their own local switches. For example, in my study, I run a single cable from the central hub, which connects to a local mini-hub that handles everything in the room (printer, computer, workstations, etc.).
- Climate Control Controller: Monitors the server room’s conditions and manages the exhaust fan. It’s part of the overall smart home system and naturally resides here, as it oversees and controls all the elements within this cabinet — lighting, fan, and temperature sensors.
- Smart Home Mini-Server: Running on a Raspberry Pi 5, this was originally hosted on the NAS. However, the NAS didn’t appreciate being constantly accessed thousands of times per second. A NAS is essentially a robust network storage device, not a full-fledged web or application server. While it can be used that way, it’s not ideal. Moving the smart home system to its own dedicated server eliminated all issues for both systems.
- Power Strip: Connected to the UPS, this serves as the primary power source for the entire setup. The UPS itself is directly wired to a dedicated electrical circuit with its own breaker. While the UPS has a few built-in outlets, they’re not convenient for plugging in bulky adapters. A heavy-duty “server-grade” power strip with extra filtering turned out to be the best solution.
- Exhaust Fan: Activates when the temperature inside the enclosure exceeds a set threshold. The controller (item 7) maintains the temperature between 80–90°F using this fan. While I could consider adding a dehumidifier, the controller’s statistics show that humidity inside the server enclosure remains very stable, so for now, it’s not a concern.
You can see exactly how much power the setup consumes and how long it lasts directly on the UPS — after all, it’s the main power source for everything:
It’s not a lot by today’s standards. I should probably add a couple more cameras around the house. They’re power-hungry and might help push the total closer to 100 watts. Right now, it’s kind of embarrassing… Calling this a “server room,” huh?
Visually, in the context of the kitchen, the server room hasn’t changed much compared to what it looked like a year and a half ago. As a result, the final project photos are nearly identical to the previous ones — aside from the spider-like router behind the glass, which adds a bit of flair (but only if the interior light is on):
For now, it’s hard to say whether this story will have a continuation. At the moment, the project is considered fully complete. Everything is set up exactly the way I wanted, everything works as it should, and I’m completely satisfied.
On the other hand, you never know when inspiration (or necessity) might strike. Right now, the only potential continuation I can foresee is if I manage to convince one of the providers to finally install fiber optics. If that happens, and if everything I’ve done here was done correctly, the process should boil down to nothing more than “pull one cable out, push the other in.” But we’ll see how things go, of course.
That’s the way it is…