dihutenosa

ffmpeg can send video over RTMP or SRT protocols. ffplay can receive and play.

https://icecast.org/faq/

In Icecast you usually have a “station” (mountpoint) sourced by a source client like IceS and streamed by Icecast to a large amount of listeners. This implies that all listeners of that mountpoint get the same stream, so an individual listener can’t pause the stream or add different tracks without affecting all other users.

You might want to check out pass - the standard Unix password manager

Surely a custom vibe-coded mess of a web frontend for ImageMagick (it allows uploading a whole lot of scanned images, then presents a comfy web UI which allows choosing particular transformations - posterize to bilevel, trim, straighten, deskew, scale-to-page-size-and-add-edges). It runs in an ancient discarded-and-salvaged laptop.

I also use the same laptop as:

• Livekit SFU (transporting media for Element (Matrix) calls)
• XMPP server (Prosody)
• STUN/TURN server (transporting media for Snikket (XMPP) calls) (coturn)
• email server (Postfix)
• hosting a QR scanner webpage, so I could scan QRs in PCs with minimal fuss (nginx + 1 static file)
• and of course, since it’s already there, home router (it’s in the basement, so too far to do WiFi. The WiFi access point runs in dumb bridge mode)

I also run a periodic HLS video broadcast (my home internet connection should survive something like 40 simultaneous watchers, I should figure out how to distribute it efficiently to other hosts). It’s too heavy for the ancient laptop, but (as it turns out, I measured yesterday), a NAS with a 4-core Intel N100 CPU just barely manages. I might switch to it, it’s probably more power efficient than the current desktop workstation.

Right you are. Still much cheaper than a laptop, let alone a server.

When logged in locally, I use btop to see an overview of what’s happening.

Other than that, I have relevant Prometheus exporters in every machine (node exporter in all machines, specific exporters by the workload), hooked up over Wireguard to my monitoring solution offsite.

The phone I actually carry around has a ntfy client talking to ntfy server on the aforementioned monitoring solution, so I get buzzes when something goes down.

Btw, does anybody happen to know where I could get a pre-cooked comprehensive alert system for my nodes? Surely many people have already written all these rules:

• if disk space > 80% consumed, send a low-priority alert
• if disk space > 95% consumed, send an urgent alert
• … everything else, there’s so much to check…

ntfy, self-hosted offsite on an old phone, it takes around 300mA, ie.less than a watt.
ntfy app on my phone.
Wireguard between them.

I found your Pimaps to be fabulous.

Haters gonna hate, there’s no help for it.

AI is a tool. Slop happens, but it doesn’t necessarily need to happen. I imagine there were many misguided, unnecessary, and just plain wrong trenches dug just after invention of the backhoe.

Go on and do even greater stuff.

Migrating the rest of my boxen to flakes rn. Kindly keep it quieter out there, folks, you’re making it hard to hear the HDD to see if the nixos-rebuild actually uses the cache.

I write my NixOS configs in my PC, commit to the repo, then push it to the server. Then SSH in, and apply it. Or more likely, MOSH in.

What’s ASD?

Absolutely amazing!

I’m doing that. Works fine. There’s an article in PMOS wiki about limiting the max battery voltage.

Syncthing? git-annex?

Manually set up Wireguard, here.

• authoritative DNS for my domain
• email (Postfix, Dovecot, rspamd)
• endlessh
• Matrix chat and video conferencing (nginx, Synapse, MAS, Livekit, coturn, Postgres)
• Prosody elsewhere for failover if international links go down
• Nextcloud
• offsite backups for important data, not the whole OS (rsync)
• QR scanner on a webpage, for transferring complex passwords locally
• Syncthing
• Offsite monitoring (Prometheus, Grafana, AlertManager) - amusingly, on an old ex-Android phone running PostmarketOS (a variant of Alpine). Eats less than a watt of power.
• Wireguard to hook up the monitoring with minimum fuss

I’d use nginx and https://git-scm.com/docs/git-http-backend

Here’s a tip: it’s one less firewall to manage, if you make your server be the router, and have your Cudy do only WiFi in dumb AP mode. Or, if the server is in a good enough location, install ´haveged´ and have it serve WiFi too.

Is there a particular reason you’re not selfhosting Wireguard, instead relying on a commercial provider? I’m running this thing on my phone, via F-Droid: https://apt.izzysoft.de/fdroid/index/apk/com.zaneschepke.wireguardautotunnel (it eats battery if you enable keepalive, but you normally don’t need keepalive, so leave it unset).