I really hope they come up with some kind of certification system for games targeting Steam consoles, in the same way Microsoft/Sony/Nintendo do. All the boring stuff like making sure controllers connect and disconnect gracefully, the console can be slept/woken at any point in gameplay without bugs, consistent language/UX etc. That stuff goes a long way to making things “just work” on a platform. IMHO it’s the one edge console still has over PC gaming. Even if it was an optional certification, it would give players some decent guidance as to what will work well.

Wow, thanks so much for the detailed rundown of your setup, I really appreciate it! That’s given me a lot to think about.

One area that took me by surprise a little bit with the HBA/SAS drive approach I’ve taken (and it sounds like you’re considering) is the power draw. I just built my new server PC (i5-8500T, 64GB RAM, Adaptec HBA + 8x 6TB 12GB SAS drives) and initial tests show on its own it idles at ~150W.

I’m fairly sure most of that is the HBA and drives, though I need to do a little more testing. That’s higher than I was expecting, especially since my entire previous setup (Synology 4-bay NAS + 4x SATA drives, external 8TB drive, Raspberry Pi, switch, Mikrotik router, UPS) idles at around 80W!

I’m wondering if it may have been overkill going for the SAS drives, and a proxmox cluster of lower spec machines might have been more efficient.

Food for thought anyway… I can tell this will be a setup I’m constantly tinkering with.

I’m curious why you feel these are easier to run on bare metal? I only ask as I’ve just built my first proxmox PC with the intent to run TrueNAS and Home Assistant OS as VMs, with 8x SAS enterprise drives on an HBA passed through to the TrueNAS VM.

Is it mostly about separation of concerns, or is there some other dragon awaiting me (aside from the power bills after I switch over)?

Floating-Point Determinism | Random ASCII - tech blog of Bruce Dawson https://randomascii.wordpress.com/2013/07/16/floating-point-determinism/

The short answer to your questions is no, but if you’re careful you can prevent indeterminism. I’ve personally ran into it encoding audio files using the Opus codec on AMD vs Intel processors (slightly different binary outputs for the exact same inputs). But if you’re able to control your dev environment from platform choice all the way down to the assembly instructions being used, you can prevent it.

The Instapak stuff I’m thinking of was basically medium sized bags that acted like a heat pack, where you break something inside the bag to combine two chemicals then shake it, which makes it expand and harden quickly.

I think the stuff is called Instapak expanding foam. Personally I think I’d remove the GPU and any mechanical drives to play it safe, but I’ve had a PC shipped to me before fully assembled with Instapak around the GPU (no HDDs, only SSDs) and it was fine. Ideally ship it in the original box for the PC case.

[Guide] NAS Killer 6.0 - DDR4 is finally cheap - Builds / [LGA1151,LGA1200] NAS Killer 6.0, Plex QSV Builds - serverbuilds.net Forums

https://forums.serverbuilds.net/t/guide-nas-killer-6-0-ddr4-is-finally-cheap/13956

Hopefully this resource is handy for you - I’m going through the same process at the moment.

Piggybacking off this, it’s worth noting if you’re adding SAS capability to your PC via one of these cards, you can look into used enterprise SAS HDDs for cheap. They’re often sold in bulk - I just picked up 72TB (12x6TB) of 7200RPM drives for AUD480 total. Availability is very region-specific and of course it’s up to you to decide if it’s worth the risk for your needs, but if you’re using RAID6 or equivalent (capable of handling two dead drives at once) the risk is minimal. Be sure to buy from sellers with a warranty (12 months minimum), and check the drives once they arrive. But in general enterprise drives are MUCH more resilient than consumer drives.

I cheated and just used a webscraper to pull the data from a local government website.

Here’s a screenshot of what my diagnostic interface for it looked like. The Python script took all the sensors as inputs and decided on a strategy for upstairs and downstairs, and gave the reason for its decision in a separate sensor for debugging purposes. The script performed all the necessary actions whenever it changed strategy.

I went down this rabbit hole a few years back. I’m in a house with 3 separate HVAC split systems over 2 floors, and we always had large temperature differentials and needed to constantly manage HVAC settings.

My goal was to never need to touch the settings for any of the HVAC units all year round, and I had plans to automate my windows so if it was nice outside, the house would ventilate automatically… but only if the pollen count was low, as I have allergies.

It became clear very early on that using the standard Home Assistant automation logic wasn’t going to be adequate for my needs, so I ended up running a Python script every half hour that reassessed the state of all sensors, using whatever crazy logic I dreamed up, then decided what actions to perform (if anything).

The general approach worked well, though I hit two snags and lost interest:

One was finding a (cheap) solution to physically open and close my roller windows. I came close, but didn’t want to make any permanent changes and had concerns about home insurance, so ultimately chickened out there. Everything still worked without this, but it would have been pretty cool to open windows automatically on a nice day for fresh air.

The other snag was more fundamental - I don’t think it’s possible to have a perfect temperature, even for one person. If I’m sitting still for long periods, I tend to want warmer temps. If I’m cleaning the house, I want cooler temps.

Ultimately my “ruleset” for the perfect solution became more and more complex, with edge cases and bugs popping up as the months rolled on, and it became clear I couldn’t represent this problem with a set of sensor inputs.

It was a fun experiment and I learned a lot, but I ended up going back to simpler automation rules after a while, like just turning on the heater in the morning if it’s cold, or vice versa. Solve the biggest problems first.

I think trying to automate to the extreme is sometimes a trap. Our hacker mindset wants the problem to be perfect and solveable, when in reality us humans are fickle creatures whose wants and whims change on the regular, and that can’t be captured easily in zeros and ones.

Two minor concerns about this approach:

1. Will the lesser known domain name make your emails more likely to be filtered as spam? I don’t know the answer, but I am fairly sure it wouldn’t help.

2. Will having your email routed through a middleman open up security issues? Probably solveable with diligence and awareness, but I recently had a non-technical friend with this setup get his Gmail breached because he was forwarding it to an email inbox on his personal domain from decades earlier that he forgot about, and didn’t have 2FA on the domain webmail. IMHO an easy oversight for anyone, honestly.

Also “Thank you for your attention to this matter.” has such “Facebook local area group” energy, like a Boomer shouting into the void about teenagers always loitering at the bus stop.

See also: https://www.abc.net.au/news/2025-08-21/cba-backtracks-on-ai-job-cuts-as-chatbot-lifts-call-volumes/105679492

The car horn on your bike is hilarious. Tangentially, a fun thing to try is using a bike bell on an escalator while standing behind people. We’re so conditioned to being dinged and moving aside that people do it before they even think about what’s going on.

No, that’s why it bought a phone.

I guess as a starting point most of us in this thread don’t really know what university research teams do.

If they had a laptop or phone, what kinds of things would they want to do that requires a server? Will they need email? Instant messaging? File sharing? Document collaboration? Will there be sensitive information? Do they need specific software? Or put another way, without this server, what can’t they do?

If you can give some hints on that kind of stuff, I’m sure people in this thread can help out more with specifics on software/tech recommendations.

Edit: obviously Unreal Tournament is non-negotiable.

I think you really need to talk to the research team to find out what they want to use the server for, and how they want to collaborate. That will inform everything else.

It’s also the fact that AAA budgets demands a huge return on investment, so you end up with gross monetization schemes designed to keep players paying as long as possible. Imagine if Stardew Valley had a Farming Pass subscription each year.

I’m sorry, what happened to the metaverse Zuck?