Yes, you read that right, hardware issues.
As you can read in my last blog post, I installed artix linux on my main server, for stability purposes, but I guess software could not fix my hardware issue. This server runs on a goofy little desktop computer that I built and converted into a server by first installing ubuntu server (I was a linux beginner at the time) and messing around with it. I then upgraded the memory, and worked on adding more services to the server. It got to the point where I could rely on it pretty well, so I took out the graphics card (yes I had a graphics card in it because sometimes I would mess up the ssh server when I was trying to configure it with something like 2fa, I was a beginner at the time please don’t crucify me) as it was using a somewhat noticable amount of power. It was only a GTX 1050 Ti with no PCIe cable, so a significantly smaller power draw than that of a modern day power-devourer behemoth card. Okay, side tangent, but do people actually use (and I mean fully utilize 90-100% capacity of) these insane wattage destroyer graphics cards??? I have a 3060 on my desktop computer and I have never used it past something around 70%. I only get it to that playing games like Half-Life: Alyx and my favorite game of all time BONEWORKS!!!!! Back on topic here: after removing the graphics card, my goofy little brain told me to remove the network card too, because I figured the motherboard’s native ethernet port is already doing everything related to networking, and the removing the little wifi network card wouldn’t do anything but help simplify the whole computer. WRONG! After I had removed the network card, the server kept shutting off all network connections after around 1-2 days. I couldn’t ssh into the server, I couldn’t even get a connection to any port on the machine. All network connections were down. For some reason, the ethernet port kept blinking in the way it normally does, indicating that it is sending packets and whatnot, but the router it was connected to claimed it was offline, and I figured it was a dhcp issue. I use dhcpcd as my dhcp client, even though I don’t really need one, as I use a static route for my server connected to that router. I would remove dhcpcd and just use iproute2 because it is minimal and I already have it installed on the system and is likely even in my initramfs, but I don’t like messing with my initramfs, as it is working, and I don’t feel like potentially breaking my system that is finally working. Back on topic here (can you tell I have ADHD?): After removing my network card, this problem kept occurring nearly every single day for several months, and I had no clue as to why it was happening. Sometimes the server would stay up for more than 2 days (!!!) but these instances happened few and far between, leaving me to literally walk over and flip the power supply switch off and on again to force restart the server (I messed up the front panel power button motherboard connector cables from the case and haven’t had the motivation to reconnect them to the motherboard lolz). For a long time I had believed that this was caused by a DDoS attack on my server, which left it unable to connect to the internet, but for some reason I didn’t think to check dmesg for any info about the missing network card. I continued trying to find software fixes to prevent DDoS attacks, but none of them worked and I thought it was just gonna be impossible to stop without being doomed to using something sucky like the unfortunate cloudflare tunnel. I thought I had finally came to my senses after all this time and realized it was a DHCP error, as I was using dhcpcd, which tried to request a dynamic ip from my router, which only gave it a static ip, which then led to it not being able to connect, as it may have had no addressed ip, but this still wasn’t the issue. I configured dhcpcd to use a static route instead of a dynamic route, but the issue still persisted. At this point, it felt as if I had exhausted nearly every possible way of defeating this problem, other than reinstalling the network card, which I thought had to be an absolutely stupid idea. I researched how many watts a network card used, and realized it only uses about 5. :| After realizing this I reinstalled it, hoping that this minor change would fix the issue, and for some reason, it did. I checked dmesg, hoping it would tell me something I could understand, but it did not and just waited it out until I could see visible results. For some reason this worked, and I’m not questioning it.
Happy new year I guess
mrgrouse out 🫡
5-31-2025 Edit: Turns out that it was either amd’s c-states that don’t work on linux and cause freezing that leaves the system basically inaccessible as it isn’t a desktop and cannot easily be sent back to a regular c-state. This, along with a faulty networking program that I’ve been using are most likely the sources of all of these problems, so I guess i was just hallucinating the rest. 🤷 Happy debugging to all of you who feel like idiots from your tech problems too :)