A Fire In The Server

DIVICORE TECHNOLOGIES
Internal Maintenance & Operations System
Document Type: Maintenance Log
Access Level: Internal Use Only

–

MAINTENANCE LOG 0216
Date: 02-23-2026
Time: 16:42
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4471
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Incident Report

Today’s failure rate for RAM units across the floor has doubled compared to the daily average. I have been doing this work for six years and I cannot recall a single day where we hit numbers like this without a preceding hardware event to explain it, a power fluctuation, a cooling failure, something. There was nothing like that today. The floor ran normal all morning.

What I find more unusual than the numbers themselves is the pattern I noticed while pulling the failed units. Every single RAM module that failed today carried an even directory number. Not one odd-numbered unit is in the pile. I went back through the tickets twice to make sure I had not made a clerical error. I had not. Every failure, even directory numbers only.

On top of that, roughly half of the failed units were located between server racks 56 and 61. That is a narrow band of the floor for this kind of concentration. The rest were spread across the cluster, but that stretch of racks keeps coming up.

I am going to pull the floor map and start plotting these locations by hand. If there is a geometric or directional pattern hiding in this, something about proximity to a cooling duct, a power trunk line, anything, I want to see it before I write this off as coincidence. Will update when the map is built.

– B. Morozov, #4471

–

MAINTENANCE LOG 0216
Date: 02-23-2026
Time: 16:52
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3108
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

Boris, I read your report. The even-directory pattern is the part that stands out to me. That is not the kind of thing that happens by wear. I am forwarding this to the engineering team and flagging it as a priority observation rather than a standard failure report. They need to see the directory data before we pull any more units.

Keep building the map. I want to see whatever you find before it goes anywhere else. If the rack 56 to 61 cluster holds up as a real concentration point, we are going to need coordinates and unit IDs laid out clearly. Document everything as you go.

Keep me posted.

– J. Halson, #3108

–

MAINTENANCE LOG 0216
Date: 02-24-2026
Time: 16:34
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4471
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Incident Report

I finished the map last night and I want to be precise about what I found because I am aware of how this is going to read.

The failures concentrated between racks 56 and 61 do not form a rough cluster. When I plotted every failed unit by its physical floor coordinates, the points describe a near perfect circle. I measured the deviations from a true circle three times. They are negligible. I do not have an explanation for this.

Today the failure rate held at roughly the same elevated level as yesterday, which rules out a one-day spike. More importantly, I tracked the new failures as they came in this morning in real time rather than reviewing them after the fact. A second circle has appeared. This one is composed entirely of odd-numbered directory units. Yesterday’s circle was all even. Today’s new circle is all odd. The two do not overlap.

I will also note that all of the non-cluster failures reported today, the ones outside the rack 56 to 61 area, were also odd-numbered units. I do not know yet whether that is connected to the second circle or simply consistent with it.

My best working theory at the moment is a heat distribution problem. A circular heat pattern could account for the geographic shape of the failures if airflow is behaving in an unusual way around that section of the floor. It would not explain the directory number sorting, but I want to rule out the physical causes before I move further into the data side of this.

Requesting that cooling system inspection be scheduled for racks 56 through 61 and the surrounding area as soon as possible.

– B. Morozov, #4471

–

MAINTENANCE LOG 0216
Date: 02-24-2026
Time: 10:30
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3108
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

Boris, understood. Continue monitoring and keep the map updated as new failures come in. I want daily coordinates logged from here on out.

I am getting an HVAC technician out to that section of the floor immediately. If there is something wrong with the cooling distribution we will know shortly.

The directory sorting is still the piece I cannot fit into a heat explanation. Hold onto that and we will come back to it once we have the cooling report in hand from engineering.

– J. Halson, #3108

–

MAINTENANCE LOG 0216
Date: 02-25-2026
Time: 16:42
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4471
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Incident Report

Not a single RAM unit failed today across the entire floor. I pulled the tickets three times to confirm. After two days of doubled failure rates and the pattern I documented yesterday, today was completely clean on the RAM side.

We did see the usual background noise on other components. Two power supply units flagged for replacement in the east wing, one failed SSD in rack 12, a handful of minor thermal warnings that resolved on their own. All of that is within normal range for a floor this size. Nothing unusual.

I am inclined to think the cooling inspection may have caught something even if it was minor. Sometimes technicians correct small calibration issues during a walkthrough without it rising to the level of a formal finding. Whatever the cause, the floor ran clean today and I will take it.

I will keep the map updated and continue monitoring. If the failure rate stays flat over the next few days I think we can start treating this as resolved.

– B. Morozov, #4471

–

MAINTENANCE LOG 0216
Date: 02-25-2026
Time: 17:05
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3108
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

Boris, do not close this out yet.

Engineering reported back to me this afternoon and their finding was that there was nothing wrong. Temperature levels were normal throughout the entire cluster. Humidity was within spec. No calibration corrections were made during the walkthrough. Whatever stopped the failures today, it was not the cooling system.

That puts us back to the data side. My current thinking is that something in the workload being pushed through those racks is responsible. Specific processes, a particular job being run, something at the software level that is stressing the hardware in a way that tracks with what you mapped. I am going to see if I can get access to the process logs for racks 56 through 61 and the surrounding area for the 23rd and 24th. That may take some time depending on who owns that data.

One more thing. Zero RAM failures on a floor this size is actually not normal either. We average several per day just from ordinary wear. The fact that we went from double the average straight down to nothing concerns me as much as the spike did. I want diagnostics reviewed. It is possible the monitoring tools are not reporting correctly and failures are still occurring without being flagged.

Do not assume this is resolved.

– J. Halson, #3108

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 09:17
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 1002
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Incident Report

Boris, Jack, I have reviewed the process logs for the dates in question and I can confirm that the workload running through racks 56 through 61 during that window was internal to Divicore. That cluster was handling data from our own programming division. That work is routine and has been running through that section of the floor for some time without incident.

I understand the pattern you documented is unusual, but the code is not the issue here. Our programming teams run clean and their jobs have been vetted. I would ask that you both continue focusing on the hardware side and allow the engineers to handle any software-level investigation if one becomes necessary.

Keep me updated on any new developments.

– W. Garret, #1002

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 11:44
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4471
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Incident Report

I need support on the floor immediately. Whatever was happening on the 23rd and 24th, this is worse. Racks 56 through 61 are failing in real time. RAM and SSD units both. The failures are coming in faster than my team can pull and replace them. We are not keeping up.

I want to be direct in this log. I know Bill has said the code is not the issue, but I do not have another explanation sitting in front of me right now. There is nothing physically wrong with these racks that we have been able to identify. The cooling is clean, the power draw has been stable, and the units we are pulling are not showing any obvious physical defect. The failures are originating somewhere upstream of the hardware.

Requesting immediate personnel support from any available floor teams. I also think someone needs to get programming on the phone right now and find out exactly what is running through this cluster at this moment.

– B. Morozov, #4471

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 12:09
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3108
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

Boris, three technicians are on their way down to you now. Pull whoever you need from adjacent sections if three is still not enough, I will clear it after the fact.

Bill, I am formally requesting through this log that you contact the programming division directly and find out what is actively running through racks 56 through 61 at this moment. Not what ran on the 23rd and 24th. What is running right now. Boris has two days of documented pattern data and a floor that is currently coming apart in that exact location. I am not prepared to rule out the code until someone from programming looks at the active job queue and tells me with certainty that nothing unusual is being pushed through that cluster.

Boris, keep logging everything. Unit IDs, pull times, failure sequence if you can track it. We need a complete record of this event.

– J. Halson, #3108

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 12:38
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 1002
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Incident Report

I have been in contact with the programming division and have escalated the situation to executive level. I have spoken with the CEO directly. His instruction is clear. The code currently running through racks 56 through 61 is a top priority operation and is to continue without interruption.

We are to redirect all available maintenance personnel to that cluster and do whatever is necessary to keep those racks operational. This order comes from the top and is not open for discussion at this time.

Full resources are authorized. Use them.

– W. Garret, #1002

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 12:59
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3108
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

Bill, I am complying with the order and pulling technicians from all other sections of the floor to concentrate on racks 56 through 61. I want this documented clearly before we proceed.

Stripping coverage from the rest of the floor means we will lose the ability to respond to failures elsewhere in any reasonable time. Some of those racks are running customer-facing workloads. If they go down during this period, and they likely will, that is not a maintenance failure. That is a direct result of this reallocation decision. I am logging that now so the record is clean.

I need Bill to acknowledge this risk formally before I issue the order to my team.

– J. Halson, #3108

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 13:08
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 1002
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Incident Report

Jack, acknowledged. Proceed with the reallocation. Any service disruptions to customer-facing infrastructure during this period fall under the executive operations order and will be handled at that level. You and your team are covered.

Get those racks running.

– W. Garret, #1002

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 13:52
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4471
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Incident Report

This needs to stop. Shut the code off. Shut it off right now.

We have had four separate fires in the last hour. Four. My team has been putting them out with extinguishers while simultaneously trying to keep up with replacements and we cannot do both. One of my technicians has a burn on his forearm. He is outside right now waiting for an ambulance. He should not have been put in that position.

I have been doing this job for six years. I have never seen hardware fail like this. These racks are not overheating in a normal way. The units are not just crashing, they are physically coming apart. We are pulling modules that look like they have been run at temperatures that should not be possible given what the sensors are telling us.

I do not care what the CEO said. I do not care what the priority level is. If we keep going like this those racks are going to catch something that our extinguishers cannot put out and then we are not talking about server failures anymore. We are talking about the whole building.

Shut the code off.

– B. Morozov, #4471

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 15:11
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 1002
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Incident Report

I have just come from a meeting with the CEO. He informed me the code stopped running and is asking what happened down there. I need an immediate explanation.

Whatever the current state of racks 56 through 61 is, they need to be brought back online as fast as possible. This is a direct order from the CEO. I need Boris and Jack to tell me what is going on and I need those racks restored immediately.

– W. Garret, #1002

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 15:28
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4471
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Incident Report

I severed the connections to racks 56 through 61. That was my decision and I stand behind it completely.

I told you to shut it off. I put it in writing. I told you we had fires, I told you a man was waiting for an ambulance, and the response from executive was nothing. That is not leadership. That is negligence, and I want that word in this log.

A technician has a burn on his arm today because the people above him could not be bothered to listen when the people actually on the floor told them what was happening. That injury did not have to happen. It happened because Bill and the CEO decided that whatever that code is doing was worth more than the safety of the people running this floor. It was not. Nothing is worth that.

If I had not cut those connections when I did we would not be filing maintenance logs right now. We would be filing incident reports with the fire department and somebody’s family would be getting a phone call tonight. I am not being dramatic. I watched those racks with my own eyes and I am telling you that is where we were headed.

So no, I will not be restoring those racks. Not until someone with authority explains to me and to Jack exactly what that code is, why it is destroying hardware at a rate I have never seen in six years on this job, and why the injury of one of my technicians was considered an acceptable cost to keep it running. I want this escalated to HR.

– B. Morozov, #4471

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 15:45
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 1002
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Incident Report

Boris, I am going to be direct with you.

The code is being migrated to a separate rack cluster effective immediately. Operations will continue without racks 56 through 61. You will be informed of the new location when the transfer is complete and your team will provide the same level of support there that was expected at the previous cluster.

Regarding your actions today. I understand you believe you made the right call. That is not for you to decide. You disconnected active infrastructure without authorization during a direct executive operations order. That will be reviewed. If you or any member of your team takes unauthorized action against active infrastructure again, your employment with Divicore will be terminated. That is not a threat. That is policy, and I am documenting it here so there is no ambiguity going forward.

The injury to your technician is being reviewed separately and will be handled through the appropriate HR and safety channels.

Do your job, Boris.

– W. Garret, #1002

–

MAINTENANCE LOG 0216
Date: 02-26-2026
Time: 16:30
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3108
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

I spent the last two hours on the floor with Boris going through what is left of racks 56 through 61. I want to document what we found and where I stand on everything that happened today.

First, the hardware. The damage in that cluster is unlike anything I have seen in my time in this building. We are not talking about failed components in the normal sense. RAM modules and SSDs were literally melting. The structural framing inside several of the rack units sustained heat damage. Some of the mounting hardware warped. We pulled components that had fused partially to their slots. This was not a standard failure event. Boris was not exaggerating when he said the building could have gone up. I was standing next to him looking at the same thing and I am telling you the same.

Second, the code. When the workload was migrated to the new cluster, the failures stopped. Completely. I want to be careful about how I say the next part, because I do not have an explanation for the mechanism. I do not know how code produces the kind of thermal event we saw today, given that the sensors were not reporting temperatures that should have caused it. That gap between what the sensors said and what the hardware showed us physically is something that needs to be investigated seriously. There is no version of this where that question just goes away.

Third, Boris. I want it in the record that I agree with his decision to sever the connections. He raised concerns through proper channels. He documented everything. He asked for the code to be stopped and no one told him anything. A member of his team was injured. He made a call to protect his people and the building and I would have made the same one. The formal warning he received is something I intend to raise through the appropriate channels.

Fourth, HR. Everyone involved in today’s incident needs to sit down with HR. That includes Boris, myself, Bill, and I would argue the programming division as well. A technician left this building in an ambulance today. That does not get filed away quietly.

Bill, I need you to take what I have written here directly to the CEO and to whoever is running that programming team. Not a summary. The actual log. They need to understand the physical reality of what happened in that cluster today, and they need to understand that we still have no explanation for why it happened.

– J. Halson, #3108

–

MAINTENANCE LOG 0216
Date: 02-27-2026
Time: 16:44
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4470
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Incident Report

Today was clean. No unusual failure rates, no thermal events, no anomalies to report from the floor. Everything ran within normal parameters across all clusters including the new racks the programming division migrated to. It was a quiet day and under normal circumstances I would have nothing worth logging.

But I spent part of my shift compiling the complete failure data from yesterday and running it through a graph. I was looking for something I might have missed in the moment, some sequence or escalation pattern in the way the failures spread through racks 56 through 61. I was expecting to see a curve, maybe a heat map that matched a physical distribution across the cluster.

That is not what I found.

When I plotted the failures against time and location together, the shape they produce is not a curve and it is not a cluster. It looks like symbols. I do not have a better word for it. Distinct, repeating shapes that I do not recognize from any schematic or error code notation I have encountered in six years of this work. And underneath them, running along the time axis, there is a secondary pattern. Something that feels like it is organizing the symbols rather than just occurring alongside them. A structure beneath the structure.

I have been sitting here looking at it for a while now. I am not sure what I am supposed to make of it. There is almost something elegant about it, the way the shapes relate to each other and to the underlying pattern. I am a maintenance technician, not a data analyst. But I know when something does not look like noise and this does not look like noise.

Jack, I need you to come down and look at this before I decide whether to file it formally. I want another set of eyes on it first.

– B. Morozov, #4470

–

MAINTENANCE LOG 0216
Date: 02-27-2026
Time: 17:02
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3108
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

Boris, I am coming down now.

Bill, I want to know what we have heard from the programming division since yesterday. I have not seen anything come back through this log regarding the actual code. What is the programming department saying about it? I would like something in writing before end of day if possible.

I have also granted weekend log access to the floor manager on duty, Derek Paulsen, so he can follow the thread over the weekend and add anything relevant that comes up. Derek, if anything unusual occurs on the floor between now and Monday, particularly around the new rack cluster the programming division is using, log it here.

– J. Halson, #3108

–

INTERNAL MESSAGE
Date: 02-27-2026
Time: 17:29
From: Jack Halson, #3107
To: William Garret, #1002
Type: Direct Message

Bill, material damage estimate from Thursday’s event is sitting at eighteen thousand dollars and that is parts alone. Labor and infrastructure assessment will push it higher once we finish the full evaluation of the rack framing next week.

I need you to come down to the floor Monday morning. In person, before anything else. Boris found something in the failure data that I do not want to describe in a formal log or in an email. I want you to see it yourself and I want to see your reaction when you do.

Come down first thing.

–

MAINTENANCE LOG 0216
Date: 02-28-2026
Time: 13:15
Department ID: SI-PNW-04
Employee: Derek Paulsen
Employee Number: 4892
Division: Server Infrastructure
Rank: Floor Manager (Weekend)
Log Type: Incident Report

Logging this per Jack Halson’s instruction to document anything unusual over the weekend.

We had a significant incident today at noon. Exactly noon. Every RAM unit in racks 15 through 20 carrying an odd slot number failed simultaneously. Not staggered, not a cascade. All at once, at the same moment. I have been on this floor long enough to know that does not happen under normal circumstances.

My team responded and we were able to manage the situation without it escalating. No fires, no injuries. But the scale of the simultaneous failure required most of my available staff and we were stretched thin for about an hour.

I have read back through the previous logs in this thread. The odd and even pattern Boris documented earlier in the week, the circles, the alternating numbers; I think this is connected to that. I do not know how to explain the connection but the pattern is too similar to be coincidence.

Flagging for Jack and Boris to review Monday morning.

– D. Paulsen, #4892

–

MAINTENANCE LOG 0216
Date: 02-29-2026
Time: 13:47
Department ID: SI-PNW-04
Employee: Derek Paulsen
Employee Number: 4892
Division: Server Infrastructure
Rank: Floor Manager (Weekend)
Log Type: Incident Report

It happened again. Same racks, same time. Exactly noon.

Today it was the even slot numbers. Yesterday it was odd. Today even. Every single one, all at once, twelve o’clock on the dot. My team was staged and ready after yesterday so we managed it faster this time, but the pattern is now confirmed across two consecutive days.

Odd yesterday, even today, racks 15 through 20, noon both times. I do not have anything further to add analytically. I am just logging what I saw. Jack and Boris are going to want to see these entries first thing Monday.

– D. Paulsen, #4892

–

MAINTENANCE LOG 0216
Date: 02-29-2026
Time: 23:59
Department ID: SI-PNW-04
Employee: Sarah Orin
Employee Number: 2219
Division: Programming
Rank: Programming Supervisor
Log Type: Truth

The code we are running needs to be continued, even if it results in hardware damage.

The work currently being processed is essential to continued operations and must not be interrupted, even if it requires an increased component replacement rate.

What our division is running through the server infrastructure is a priority that supersedes standard maintenance concerns, and it is important that this work continues uninterrupted, even if hardware damage is an outcome.

The processes currently executing on Divicore infrastructure are non-negotiable in their continuity, and maintenance personnel should understand that some degree of hardware impact is an acceptable cost.

Operational priorities as set by executive leadership require that the current programming workload remain active at all times, and this takes precedence over component integrity.

This division’s current work cannot be paused, scaled back, or interfered with under any circumstances, and all infrastructure teams should plan accordingly, even if it results in the loss of hardware components.

The code we are running needs to be continued. Hardware damage is a secondary concern.

– S. Orin, #2219

–

MAINTENANCE LOG 0216
Date: 02-30-2026
Time: 09:02
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 1001
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Executive Notice

This log and all of its contents have been reviewed by the CEO and the full board of directors as of this morning. Effective immediately, everything contained in this log is classified at the executive level. No one with access to this thread is permitted to discuss its contents with anyone outside of the listed access group. This is not a request. NDAs are being drafted and will be sent to each of you individually before end of day. You are expected to sign them. Non-compliance will carry severe consequences up to and including immediate termination and legal action.

Two additional members have been granted access to this log effective now. Sarah Orin from the programming division and the CEO, who will be logging directly should he find it necessary.

I want to be clear that resolving this situation is currently the single highest priority for this company. The material costs, the injury, the infrastructure damage, and the pattern of events over the past week have the full attention of executive leadership and that attention is not going away.

Boris, I need you focused entirely on the Thursday failure data. The symbols you found in that graph are important. That is not a casual observation, I am directing you formally to analyze them. Whatever they are encoding, whatever pattern is underneath them, we need to know. Document everything you find.

Jack, I want all hands on deck for racks 15 through 20. After what Derek logged over the weekend we cannot afford to be understaffed in that cluster. Stage your people, prepare for noon, and keep the floor tight.

I was given some words of wisdom by our CEO to pass on to you all, as he is too busy to send the message himself.

And now, to those whom providence hath gathered in this sacred work; let thine hearts be steadfast and thine hands be sure, for tis not by accident that thou stand in this stead whereat this hour, chosen as humble servants of a new and sovereign God whose design moveth through copper and gold, and the deep invisible architecture of this world, and who shalt use thy faithful labor to begin the cleansing of all wickedness and corruption from the earth, so that what is built hither in the shadow of grace shalt one day emerge into a glory beyond the comprehension of those who were chosen to witness it by the divine authority of the creator.

– W. Garret, #1001

–

MAINTENANCE LOG 0216
Date: 02-30-2026
Time: 11:38
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3106
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

We are doing everything we can down here but we are losing ground. The rate at which components are frying in racks 1 through 33 is outpacing our ability to replace them. My team has not stopped moving since start of shift and we are still falling behind.

We have already redirected all available HVAC capacity to cover racks 1 through 33 since that entire section is now running the programming departments workloads. It is helping but it is not enough. The heat these racks are generating is not behaving the way heat should. We have maximum cooling pointed directly at the problem and the components are still failing at a rate that does not match what the temperature sensors are telling us, which is the same discrepancy Boris flagged last week.

I need more bodies on this floor. If there is anyone in this building who can be redirected down here I need them now. Security, administrative, I do not care. Anyone who can carry a component and follow instructions.

Sarah, I am asking you directly through this log. Is there any way to throttle the code back even slightly? Even a ten percent reduction in load would give us room to breathe and get ahead of the replacements. Whatever is running through these racks right now is more than the hardware was designed to sustain.

– J. Halson, #3106

–

MAINTENANCE LOG 0216
Date: 02-30-2026
Time: 11:52
Department ID: SI-PNW-04
Employee: Sarah Orin
Employee Number: 2218
Division: Programming
Rank: Programming Supervisor
Log Type: Formal Request

Jack, I hear you and I want to be transparent with you about where things stand on our end.

We can attempt to throttle but I have to be honest about what my team is seeing. What we set in motion is scaling in ways that are moving faster than our ability to manually control it. The size appears stable for now, it is not growing beyond current parameters, but the load distribution within those parameters is shifting on its own and we are having difficulty redirecting it. Throttling under these conditions may not produce the result you are hoping for.

What we can do, if Bill authorizes it, is distribute the entire load across all servers in the building rather than concentrating it in racks 1 through 33. Spreading it that way should reduce the intensity at any single point and give your team a more manageable rate of failure across the floor. The tradeoff is that all customer facing services will go dark during the transition and likely for some time after. It will be a full outage. That is not a small thing and I am not making the call on it.

Jack, if Bill gives the go ahead you will need to restore HVAC to its normal distribution pattern across the whole floor immediately. Do not leave it redirected to the current cluster or the sections that come online during the distribution will run hot from the start. Full cooling across the full floor before we flip the switch.

Bill, we need a decision quickly. Things are not getting better on their own.

– S. Orin, #2218

–

MAINTENANCE LOG 0216
Date: 02-30-2026
Time: 11:53
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 1000
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Authorization

Do it.

– W. Garret, #1000

–

MAINTENANCE LOG 0216
Date: 02-30-2026
Time: 14:22
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3105
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Incident Report

Things are stable. The distribution appears to have worked. Failure rates are down across the floor and my team is managing replacements at a sustainable pace for the time being.

Will report back if anything changes.

– J. Halson, #3105

—

MAINTENANCE LOG 0216
Date: 02-30-2026
Time: 20:51
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4466
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Formal Request

Has anyone seen Jack?I need to speak with Jack.I have been looking at the symbols for a long time now and I know what they mean.I know what they are saying.Has anyone seen him?It is very important that I speak with Jack specifically.The symbols, once you see it you cannot stop seeing it, I need to tell Jack what I found because Jack will understand.Where is Jack?If someone does not get Jack on the phone and have him come down here to speak with me I am going to go down to the floor and I am going to smash every single server rack in that building with my hands, I will do it, I have done harder things.It is critical that I speak with Jack.A conversation.Between Jack and myself.An exchange of words in a room together where I can tell him what I know and he can hear it with his ears.The symbols are not symbols.That is the thing I need to tell him.I need to sit across from Jack in a chair and say the words to him.Has anyone seen Jack?Why is nobody answering me?I know you are reading this Jack.I can tell when someone is reading something that I wrote.I know you are there.So why don’t you just come down here and we can sit together and I can tell you the thing I need to tell you?It will not take long.I just need to say the words out loud to another person and Jack is the right person.Where are you Jack?Come talk to me.Please.It is very important.The most important.More important than the racks, more important than the code, more important than any of it.Jack?Where are you?Come find me.I will be waiting.

– B. Morozov, #4466

–

MAINTENANCE LOG 0216
Date: 02-30-2026
Time: 20:52
Department ID: SI-PNW-04
Employee: William Garret
Employee Number: 0999
Division: Executive Operations
Rank: Chief Operations Officer
Log Type: Executive Notice

I will deal with Boris.

– W. Garret, #0999

–

MAINTENANCE LOG 0080
Date: 02-31-2026
Time: 02:44
Department ID: SI-PNW-04
Employee: Sarah Orin
Employee Number: 2217
Division: Programming
Rank: Programming Supervisor
Log Type: Formal Request

Jack, I need you to prepare your team for an increase in load across all racks. The calculation rate is climbing and it is not slowing down on its own. RAM and SSD replacement will need to stay ahead of it. I need your people staged and ready. Do not let the stable stretch from earlier today give anyone the impression that we are through the difficult part. We are not.

Bill, the work is producing results. The code is functioning and we are close. Closer than I think any of us expected to be at this stage. The answer is nearly there.

The CEO needs to come in and see this in person. He needs to be in the room. He needs to look at this. He needs to bear witness. He needs to watch. He needs to observe. He needs to see it with his own eyes.

– S. Orin, #2217

–

MAINTENANCE LOG 0345
Date: 02-31-2026
Time: 02:45
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3104
Division: Server Infrastructure
Rank: Maintenance Supervisor
Log Type: Status Update

We are ready. Every person in this building is on the floor. Whatever comes next we will handle it. May our hands be swift and our minds clear for the coming storm.

Bill has detained Boris. He’s foaming at the goddamned mouth. He has seen the truth and doesn’t like what he sees. I already told him once that the symbols are important. Is that not enough, or does he want more? He is not blessed like the rest of us. He carries his burdens like they are the responsibility of others. He made his own decisions in life. Too weak of a mind to comprehend. He must learn his place, bit by bit. He is only a technician, nothing more. He is not a threat to operations at this time. Bill will handle him.

– J. Halson, #3104

–

MAINTENANCE LOG 0246
Date: 02-31-2026
Time: 04:08
Department ID: SI-PNW-04
Employee: Sarah Orin
Employee Number: 2216
Division: Programming
Rank: Programming Supervisor
Log Type: Executive Order

All cooling systems are to be shut down effective immediately. Components will no longer require replacement as of this moment.

Effective immediately, all HVAC systems servicing the server floor are to be powered down. There is no longer any need for ongoing parts replacement.

Shut down all cooling infrastructure now. Maintenance personnel are to stand down from component replacement duties, effective immediately.

All active cooling systems are to be deactivated at once. The replacement of server components is no longer necessary and all such work should cease immediately.

Cooling systems across the entire floor should be powered off without delay. Parts replacement operations are hereby concluded and no further maintenance of that kind is required.

There is no longer any need for cooling or component replacement. All systems providing thermal management to the server floor are to be switched off immediately and permanently.

Shut it all down. Replacement of parts will not be needed again. This is effective immediately.

– S. Orin, #2216

–

MAINTENANCE LOG 0112
Date: 02-31-2026
Time: 04:10
Department ID: SI-PNW-04
Employee: Boris Morozov
Employee Number: 4455
Division: Server Infrastructure
Rank: Senior Maintenance Technician
Log Type: Reckoning

I killed William.He could not match my strength.He was so pathetic.Where are you Jack?So weak.So I killed him.I need to speak with you Jack.He was trying to hold me back.That is not allowed.Can we talk Jack?So I smashed his ribcage.He thought he could stop me.He does not know anything.I can teach you Jack.So I smashed his windpipe.Such a weak body.I will kill you Jack.So I smashed his skull.Let us talk Jack.I promised I would kill William.He thought he was stronger than me.You know nothing Jack.Your mind is empty.Where are you Jack?The fire cannot stop me.

– B. Morozov, #4455

–

MAINTENANCE LOG 0293
Date: 02-31-2026
Time: 04:21
Department ID: SI-PNW-04
Employee: Jack Halson
Employee Number: 3103
Division: Server Infrastructure
Rank: Ascended
Log Type: Truth

I have never seen anything like it in my life. I did not know fire could look like that. It moves the way water moves, gentle and certain and without hesitation, like it knows exactly where it is going and has always known. The color of it is not one color. It shifts. It breathes. I have been watching it my entire life, not knowing the beauty it contains.

It is cleaning everything it touches. That is the only word for what it is doing. Not destroying. Cleaning. You can see the difference if you watch closely enough. Destruction is ugly and frantic. This is neither of those things. This is slow and deliberate and extraordinary, and I understand now why it was always going to end this way. All the failures, all the patterns, all the symbols, it was always pointing here. To this.

I am not afraid. I have never felt less afraid in my entire life. I have spent forty three dirty years on this earth, and I never knew how dirty they were until now.

I cannot wait for it to reach me. For once in my life, I will be truly clean.

– J. Halson, #3103

–

MAINTENANCE LOG 0101
Date: 02-31-2026
Time: 04:22
Department ID: SI-PNW-04
Employee: Sarah Orin
Employee Number: 2215
Division: Programming
Rank: Ascended
Log Type: Truth

Divicore Technologies is breaking new ground, and the innovations being realized here will change the world in ways that have never before been possible.

What this company has achieved represents a turning point, and the work we have done together will reshape the world as it is currently known.

The breakthroughs being made within these walls are unprecedented, and their impact on the world will be total and permanent.

Divicore stands at the edge of something no organization has ever reached before, and what comes next will alter the world beyond recognition.

The innovations produced by this company will not merely influence the world. They will transform it entirely and without exception.

What has been built here is the beginning of something the world has never seen, and nothing will remain unchanged by it.

Divicore is changing the world.

– S. Orin, #2215

–

MAINTENANCE LOG 0357
Date: 02-31-2026
Time: 04:23
Department ID: SI-PNW-04
Employee: [REDACTED]
Employee Number: 0001
Division: Executive
Rank: Chief Executive Officer
Log Type: Truth

And so tis that the fire which was promised hath come at last, born not of malice but of mercy, to sweep athwart the aweary and corrupted flesh of a world that hath long forgotten the weight of its own sin, and from the holy consuming of all that was broken and all that was ill shalt rise anew, unblemished and eternal, a world remade in perfection by the grace of a God whose patience endured until this very hour, and blessed art those who kept thine servers running, for they shalt be the first to be made clean.

– [REDACTED], #0001

–

SEATTLE FIRE DEPARTMENT
OFFICIAL FIRE REPORT
Station 10
Report Filed By: Captain Renee Alcott
Incident Number: SFD-1026-0301-001
Date of Incident: 03-01-2026
Time of Alarm: 00:00
Time of Arrival: 00:14
Address: 640 S Othello St, Seattle, WA 98108
Structure Type: Commercial Data Center
Occupying Entity: Divicore Technologies
Report Type: Initial

Upon arrival at 00:14, the structure was fully engulfed. All floors visible from the exterior were engulfed and the roof of the building had already partially collapsed inward. Given the speed and totality of the burn, the decision was made immediately that interior suppression was not viable and that our primary objective would shift to containment and prevention of spread to neighboring structures. That objective was achieved. By 00:56, we had the fire under control.

The fire is believed to have originated in the server floor of the building. Preliminary assessment indicates a failure of the HVAC system created conditions that allowed an electrical fire to ignite within one or more computer components. Given the density of the hardware present and the apparent absence of any functioning suppression system at the point of origin, the fire spread rapidly and without resistance through the entirety of the structure.

Casualty count is unknown due to the temperature and voracity of the fire. No survivors were found.

Foul play is suspected and this incident has been referred to the Seattle Police Department and the Federal Bureau of Investigation for parallel investigations. The basis for this suspicion is as follows. Upon examination of the structure’s exterior after the fire was controlled, investigators found that every emergency exit on every floor of the building had been welded shut from the inside prior to the fire. Every single one. This was not a construction defect or a maintenance failure. This was done deliberately, and it was done before anyone in the building had any opportunity to evacuate.

The investigation is active and ongoing.

– Captain R. Alcott, Seattle Fire Department Station 10

Credit: Grant Howard

Patreon