The most straight-forward method would be to buy a standalone switch. I have a TP-LINK TL-SG108 8-port gigabit switch and it seems to retain the ARP table indefinitely.

My previous solution was an ESP32 board with an SSH server and a relay, wired parallel with the power switch, that would be closed by an output pin on command.

How much experience do you have with networking, exactly?

The DNS record points to a private IPv4 address (10.0.0.41), which cannot be accessed from the internet for multiple reasons; first of which is that it’s almost certainly behind a NAT gateway.

Your internet provider has given you a single publicly routable IPv4 address and assigned it to the WAN interface on your modem or router. If you want to access a host on the LAN, you’ll first have to configure port mapping or port forwarding on the router. Then you’ll have to open holes in your firewall and accept the fact that every bad actor will try to break into that host unless you know how to set up network security.

Better out-of-the-box hardware support, in my experience. We have a machine learning server at work, it didn’t see the GPUs on Debian Bullseye with the driver versions specified by the manufacturer, but worked perfectly with Ubuntu Server out of the box.

A distribution that is preconfigured by professionals has great value in a practical setting, even if that value has diminished in the eyes of the kind of person that Lemmy attracts. If I had tried to get Debian working by overruling the manufacturer’s instructions, I’d have to take responsibility for it, both its maintenance and the downtime and potential damage if I had fucked something up. With Ubuntu, I get to delegate at least part of the responsibility to Canonical (while covering my own ass), and that’s something you can’t backport.

It’s not about saving ten seconds of typing. It’s about supporting OP’s claim with credible sources. And I have to stress credible because mainstream journalism has a tendency to turn unverified or downright false claims into a woozle.

Linux has two different kinds of “used” memory. One is memory allocated for/by running processes that cannot be reclaimed or reallocated to another process. This memory is unavailable. The other kind is memory used for caching (ZFS, write-back cache, etc) that can be reclaimed and allocated for other things as needed. Memory that is not allocated in any way is free. Memory that is either free or allocated to cache is available.

It looks like htop only shows unavailable memory as “used”, while proxmox shows the sum of unavailable and cached memory. Proxmox “uses” 11 GB, but it’s not running out of memory because most of it is “available”.

In his newest (and worst) How Do You Do Fellow Kids moment, Mark Zuckerberg launches the Poob service, accessible exclusively through the Metaverse. What does it do? Fucked if we know.

adduser is an interactive wrapper for useradd. It can, for example, prompt the user to set a password rather than execute passwd separately. Very useful if you just want to manage a user without reading through useradd’s command line options, then running usermod because you forgot to set something.

It doesn’t excuse the bad naming, I’d rather have something like useradd --interactive, but it’s worth remembering.

I wish my one bad experience in 2015 had been the absolute last time Pulse failed for anyone ever. Alas, time doesn’t work that way, and Pulse remained failure-prone for years after my encounter with it.

Extremely frequently. Digital musical instruments generally don’t output production quality sounds – they output MIDI data that describes what note is being played, and an audio synthesizer device or software interprets it and generates the audio data.

Very mature. If you’re going to act like a toddler, I’ll put you in the timeout box.

For the record, your reply to my comment had no bearing on my decision, and I’ll preserve the entire comment chain in case I find myself on YPTB.

Very mature. If you’re going to act like a toddler, I’ll put you in the timeout box.

Memes are not necessarily meant to be funny. They’re meant to represent a common experience within a group. The horrid failure rate of PulseAudio just a couple years ago fits that definition perfectly, and in fact it’s why I gave up on Ubuntu 14.04 in 2015.

It should be Neovim and Lua. Nobody should be subjected to the curse and torment of writing Elisp.

I think you’re confusing it with Manjaro, which has had several.

Proxmox is a great starting point. I use it in my home server and at work. It’s built on Debian, with a web interface to manage your virtual machines and containers, the virtual network (trivial unless you need advanced features), virtual disks, and installer images. There are advanced options like clustering and high availability, but you really don’t have to interact with those unless you need them.

Not even the man RMS himself uses GNU/Hurd or Guix, which is hilarious.

You should check out the tldr program. It’s a community-driven quick reference tool that lists common practical examples for commands.

Windows devs used to have fun.

New developments: just a few hours before I post this comment, The Register posted an article about AI crawler traffic. https://www.theregister.com/2025/08/21/ai_crawler_traffic/

Anubis’ developer was interviewed and they posted the responses on their website: https://xeiaso.net/notes/2025/el-reg-responses/

In particular:

Fastly’s claims that 80% of bot traffic is now AI crawlers

In some cases for open source projects, we’ve seen upwards of 95% of traffic being AI crawlers. For one, deploying Anubis almost instantly caused server load to crater by so much that it made them think they accidentally took their site offline. One of my customers had their power bills drop by a significant fraction after deploying Anubis. It’s nuts.

So, yeah. If we believe Xe, OOP’s article is complete hogwash.