A software developer and Linux nerd, living in Germany. I’m usually a chill dude but my online persona doesn’t always reflect my true personality. Take what I say with a grain of salt, I usually try to be nice and give good advice, though.

I’m into Free Software, selfhosting, microcontrollers and electronics, freedom, privacy and the usual stuff. And a few select other random things, too.

  • 1 Post
  • 129 Comments
Joined 5 months ago
cake
Cake day: June 25th, 2024

help-circle

  • Well if you want a proper upgrade, 40TB plus redundancy and space for a GPU, I’d say you don’t want a mimi PC but a full-blown one. I built my server myself from components. It’s hard to find good numbers on power consumption and that was one of my main concerns. I had a look at some PC magazines and what kind of mainboards they recommend for a home server. Figured I wanted 6 SATA ports and I started from that. Unfortunately said magazine doesn’t have a good article right now, so I don’t know what to recommend. Another way is to look for refurbished PCs. If they’re some brand like Lenovo or Dell, you’ll find the specs online. With a N100 mini pc, I’m not so sure if that’s a big step up from your current setup… I don’t think they have more internal harddrive ports or slots for GPUs than your current laptop.


  • Very good answer. I’ve also spent some time analyzing some red herrings when it was something else like a bad cable or connector. And by the way, you can use the same keys in journalctl as in the usual pager (less(?)) so hit / and search for ‘unmount’, ‘disconnect’, etc. And then scroll through the log and find out what led to the situation.








  • Check out yunohost.org (and similar projects) If you’re in for a turnkey-solution.

    But yes, a reverse proxy that does all the work and handles SSL is a nice solution. I also use that. It’s relatively easy to set up, doesn’t really slow down anything and makes a lot of stuff easier to manage.

    I use NGinx, but Caddy or Traefik will do the same. And I don’t use Cloudflare, so I can’t comment on that.

    And btw, Jitsi-Meet is going to require some more dedidated ports for the WebRTC, STUN, etc











  • Correct answer. There is no general purpose AI model that can fit ino 1GB. These small models exist, but they do very specific small tasks. Sentiment analysis, object detection, word embeddings for vector databases…

    For coding, answering questions and generating text, you’d need like 6-8GB minimum. For maths way more than that and they’ll still be throwing dice instead of giving correct answers.