Makes sense. Your 2nd definition is what I take from the term scaling. Let’s see if op comes back with any notes.

For video encoding, I run an 8th gen Intel i5 8500t. The quicksync is good enough for nearly anything 1080p.

Not sure what you mean by the “scaling”.

Nice work. I would read more articles like these.

“Pleb” is generally used as a pejorative and is roughly equivalent to calling someone a peasant.

Are you not using LE certbot to handle renewals? I can’t even imagine doing this manually.

I don’t know… I’m skeptical of its bandwidth.

I love alpine, and I use it where I can. And it has many advantages over other distros and setups. But a declarative, ram-only distro that boots over the network doesn’t help manage non-conformant machines.

I still need to manage Debian, old centos boxes, Ubuntu machines, and a couple old-as-time sun machines. Nixos isn’t the tool for that job. Ansible has two dependencies: ssh and python, and there are ways around the 2nd one. Ansible works really well here.

Not trying to bash nixos, here, but I’m not sure why so many users on Lemmy compare ansible and nix, they don’t really operate in the same spaces.

Can you manage a Debian server with nix?

Hahaha nice. I have PTSD from teaching the interns how to search graylog and then locking up elastic with piped conditional searches.

I appreciate the response.

I’m still keeping an eye on these technologies and I hope we can set some decent standards for alternates to WiFi and LTE narrow band.

I’ve been watching halow for a while, I haven’t yet seen any sustainable, real-world examples beyond a few hundred kbps (not bytes). I have seen the 1Mbps results, and they’re promising, but most places with any other traffic in the free band is busy. If you have any successful and repeatable tests hitting at usable speeds, I’d love to see them.

After getting into meshtastic and a few other lorawan projects, I’m a bit concerned that tests for these are always high and visible, which doesn’t work well in the mountains, even at shorter ranges.

I used to be more hands-on with these new standards, but I’ll wait for better tests to come from halow before I try it out.

Definitely possible, and there is already some tech with the kinks worked out like wimax that could wirelessly serve a whole town. There are also folks who have created their own isps to fill in where the big players don’t bother. It is apparently regulatory hell to get up and running.

The problem isn’t technology, it’s people.

Like every protocol in the unlicensed 900mhz range, 802.11ah has a very limited transmission rate in the 50 to 100 kbps range.good for occasional data like sensors or a few bytes of message, but not for any modern comma like AV, mass file transfer, etc.

If you’re not fond of manipulating config files manually, just use nmcli (from your link):

You can get an idea of NetworkManager’s settings by running nmcli on the command line.

It is a bunch simpler. The days of just raw-dogging resolve.conf and nsswitch are long behind us.

Aren’t these docs an admission that it’s a clusterfuck?

The Debian wiki admittedly needs work, but it is a wiki, so make an account and update what you think is lacking or unclear.

Anubis is an elegant solution to the ai bot scraper issue, I just wish the solution to everything wasn’t just spending compute everywhere. In a world where we need to rethink our energy consumption and generation, even on clients, this is a stupid use of computing power.

You are asking the WG server to listen to incoming requests from outside your lan subnet, so it is ignoring VPN requests from that subnet.

There are two solutions to this:

1. Add routing to your wireguard server instance to allow the VPN intermediary subnet to accept connections from your lan subnet or
2. Allow your wireguard client to split-tunnel, so it can reach subnets that aren’t reachable outside your WG tunnel.

Hey, I get it. It’s really difficult these days to tell when to care and when caring will make living in the modern world too difficult.

And I will uphold your right to care about different stuff than I do.

OK, thanks.

Not really that helpful, it’s just more he said/she said.

Grapheneos is a pretty poor choice of example, it has its own issues, they have conflicts with many others (see France pullout), they have openly fought with Murena, e/os, fairphone, among others. Plus, grapheneos likes to throw shade on “less permissive licenses”, which is really weird, considering they use an MIT license, whose loose terms is abused by many not-so-great companies.

The lead of grapheneos is just as controversial as futo, gets into public spats with others, generally difficult to work with, etc.

I say this using GrapheneOS myself.

I’m not a fan of the Curtis Yarvin association, though.

The people running these companies aren’t perfect. Sometimes they aren’t ideal, either. But I’d rather use Immich than Google photos and heliboard than Swiftkey.

Is there any other info on this besides drew devault’s blog post?

I “document” everything by forcing myself to create ansible runbooks for new services and configs. I have some gaps, definitely, but the more of them I create, the easier new services are to deploy.