Hell yeah! 10x speed improvement for free!

What I’m noticing more, is that you can keep a consistent 11.4MB/s, this feels relatively close to what you’d usually pull through a 100mbit/s link (after accounting for overhead). If that’s the case, it shouldn’t matter how the NFS client decides to chunk the data, for how much throughput there is to the NAS. Which means you’re looking at a broken NFS server that can’t handle large single transmissions.

If it’s not the case, and you’ve got a faster network link, it seems that the NAS just can’t keep up when given >2gb at once. That could be a hardware resource limitation, where this fix is probably the best you can do without upgrading hardware. If it’s not a resource limitation, then the NFS server is misbehaving when sent large chunks of data.

Basically, if your network itself (like switches, cables) isn’t broken, you’re either dealing with a NAS that is severely underspecced for what it’s supposed to do, or a broken NFS server.

Another possibility for network issues, is that your proxmox thinks it has gigabit (or higher), but some device or cable in between your server and NAS limits speed to 100mbit/s. I think it’d be likely to cause the specific issues you’re seeing, and something like mixed cable speeds would explain why the issue is so uncommon/hard to find. The smaller buffers more frequent acknowledgements would sidestep this.

Do note I am also not an expert in NFS, I’m mostly going off experience with the “fuck around and find out” method.

Sounds like a band-aid fix to a completely different problem. If NFS is timing out, something is clearly broken. Assuming it’s not your network (though it could very well be), it’s likely the Synology NAS. Since they’re relatively closed devices afaik, I sadly can’t help much in troubleshooting. And sure, dumping 25GB on it all at once is heavy, but it should handle that, being a NAS.

Iirc, the XZ backdoor was specifically targeting RH and Debian, which for some reason link libsystemd into OpenSSH. Afaik even upstream Arch was unaffected, not just Artix. The exploit code, though non-functional, still made its way onto your system (assuming you updated when it was in a release version).

I’m not defending systemd though, it’s clear that Poettering’s goals do not align with the rest of the Linux community. I’m saying that Artix not being affected by the XZ backdoor is not a good argument for why to use Artix or avoid systemd.

It’s like saying “Linux doesn’t get malware” because most desktop malware targets the OS with the largest desktop userbase, Windows. This alone doesn’t suddenly make Linux “better”. That doesn’t mean there aren’t other reasons to avoid Windows.

The smaller/newer distros have no evidence of staying around for years, so it’s hard to judge whether they’ll be around in another couple years. Distros like Bazzite are definitely interesting, but you can’t reliably predict whether it’ll get updates in 10 years. There are stable community-led distros that have been around for a long time, like Debian.

Personally, I have nothing against crawlers and bots

If they’re implemented reasonably, web crawlers aren’t the issue. The problems with them mostly stem from laziness and cost cutting. Web crawlers by AI comapnies frequently DDoS entire services, especially Git forges like Gitlab or Forgejo. Not “intentionally”, but because these crawlers will blindly request every URL on a service, no matter the actual content. This is cheaper for the AI company to implement this way, and scan through the data later. But this also leads to the service having to render and serve tens of thousands of times as much content as is actually present. They are made to try and hide themselves doing so, which is the biggest reason we see “modern” PoW CAPTCHAs everywhere, like Anubis or go-away.

Robots.txt used to work, because search engines needed there to be an “internet” to provide their services. Web crawlers pre-AI were made knowing that taking down a service made another website go down, which lessened the overall quality of search results.

I’ve had LLM webcrawlers take down my whole server by DDoSing it several times. Pre-LLMs, a git forge would take maybe a couple hundred MB of RAM and be mostly idle while not in use. Nowadays, without a PoW CAPTCHA in front, there are often over 10.000 active concurrent connections to a small, single person Git forge. This makes hosting costs go through the roof for any smaller entity.

Is the video player application itself muted in pipewire? What’s the output device set to?

You can check these things with an application like pavucontrol. Pipewire (and pulse) have a default audio device, but individual applications can set a different audio device if they want to.

Another great category of utilities for pipewire is virtual patchbays. If you’re looking for something simple, helvum or qpwgraph are geat. For all the technical details in a GUI, coppwr provides a good experience.

Absolutely true, but there’s one clear and obvious way; drop support for the project yourself.

If a FOSS project is archived/unmaintained, for a large enough project, someone else will pick up where the original left off.

FOSS maintainers don’t owe anyone anything. What some developers do is amazing and I want them to keep developing and maintaining their projects, but I don’t fault them for quitting if they do.

It’s still made by the slop machine, the same one that could only be created by stealing every human made artwork that’s ever been published. (And this is not “just one company”, every LLM has this issue.)

Not only that, the companies building massive datacenters are taking valuable resources from people just trying to live.

If the developer isn’t able to keep up, they should look for (co-)maintainers. Not turn to the greedy megacorps.

The exact numbers for when it messes something up, but keeps running, are unknown and highly ubpredictable.

According to above post, about 10% of firefox crashes (more numbers found in the post) are caused by this stuff. It’s not unreasonable to say those crashes could’ve had the bitflip happen on content instead, changing maybe a character on the page or something.

Note that it’s not 10% of users, as that’s reslly hard to figure out. Someone with bad RAM will likely crash more often.

Simple stuff like a calculator can be just as broken by a bitflip as more complex things. You wouldn’t want your calculator to say 1 + 1 = 2049.

If you want to rely on your computer, ECC RAM is required.

Arch isn’t exactly intended for beginners, and the “windows is more bloated than Linux” thing applies for pretty much every Linux distro.

If you’re willing to spend a significant amount of time learning and tinkering, a distro like Arch could provide what you need. However, if you’re just starting out, you might want to start with something easier. Distros like Linux Mint or Fedora are plenty lightweight compared to Windows, and they require much less learning to get started actually using your computer.

As for “which distro allows you to use the cheapest PC”, this mainly comes down to how much effort you’re willing to put in, not necessarily the distro you use. At a certain point, a desktop computer will “just work” for basic desktop tasks, the distro doesn’t change much there.

If I had to pick one, Arch. I already use it a lot, so it’s familiar. I know my way around the package manager and how to create packages, so even when things aren’t available for Arch out of the box, I can make it work.

It’d be kind of a hassle trying to keep anywhere close to 100% server uptime, but for my own personal stuff that shouldn’t be that big of an issue, as I can fix it when I have the time.

For desktop, I basically can’t do stable release. I frequently mess with new projects requiring the latest versions of everything, which is a near impossible task on stable-release distros.

It’s the task of your CPU scheduler to ensure your system doensn’t freeze, even over 100% CPU usage. If it’s completely unresponsive, it’s more likely you’re running out of memory instead.

In my experience, the Debian installer is just confusing. Once you’re past that, the userbase is smaller than Ubuntu’s. Their repos are different too, meaning software packaged for Ubuntu isn’t guaranteed to work on Debian. Ubuntu itself is pretty terrible for its own reasons, so when asked for a desktop Linux distribution “close to Ubuntu” I’d put Mint first. (For general recommendation, I’d probably say Fedora now.)

Debian 13 is still relatively new, so the problems of it being out of date aren’t showing yet. Debian 12 just before 13 released had tons of these issues, like glibc being too old for some binary programs, or the kernel not being new enough for some “gaming” features.

For reference, I am on Arch Linux. I feel I have a good understanding of how to manually install Linux. The Debian installer confused me in many ways, the main one being that “language and region” are closely tied, and selecting en_US “language” forces you to choose an American timezone later in the installer. In general it was a slow install process too. This is something other “user friendly” distros handle much better. A default live environment, a quick installation, and options being there, but having the defaults automatically correct (like timezone).

Like (almost) every other distro, Debian has its own benefits and downsides. These make it a good fit on desktop for slightly more experienced users, or users familiar with apt. This means it isn’t in the list of distros I’d generally recommend to people when they’re not familiar with Linux.

Matrix (Synapse with Element) can be self-hosted for free, though they have optional paid plans for enterprises. The main goal of Matrix is federation (connecting with other servers), though this can be turned off completely. This is probably the most “business” look/feel you can get fully FOSS, if that’s what you’re looking for.

XMPP has more clients/servers, and is more for the technically oriented end user. I can’t really give recommendations here, as I haven’t extensively used XMPP.

Spacebar (formerly Fosscord) is a Discord clone (API compatibility as a goal) that can be selfhosted.

I use whatsapp for communicating with a few groups of people who are refusing to switch over. I’m not getting them to move, they’re a mix of tech illiterate, ignorant, or just unwilling. These groups have important annoucements in them, that I’m unable to receive any other way. Sometimes there’s just no way around services like these.

I do my best to protect myself (no proprietary facebook code ever runs on my main device), and keep personal information I provide to a minimum (as I don’t trust Facebook’s E2EE claims). There’s not much more I can do besides not receiving important announcements.

CachyOS has a structure that’s much closer to Arch than Manjaro, but they still replace the majority of Arch repos with their own. I’ve seen both their repos and the optimizations they apply being the root cause of issues on some friends installs.

Use whatever you’re comfortable with, and what you know works.

On that note, Manjaro and CachyOS don’t work. You should avoid them. They both make changes that harm reliability, and both frequently make avoidable mistakes (especially Manjaro). If you need something like those two, EndeavourOS is a better option, or just base Arch Linux.

Arch Linux itself is a good distro, but made for a specific target audience. If you want to tinker with your system and learn along the way, it may be a good option for you. If you want to “set and forget” your media center PC, a stable-release distro like Debian or Rocky might be better options for you.

I don’t think this has happened yet with video games, however it is in no way illegal for Valve to do this. There’s been plenty of examples of other media being ripped away from consumers, like “purchased” movies and music.

On Steam, you are purchasing a license to play a game, not the game itself. At any point and for any reason, Valve can legally revoke this license or restrict access to it.