• 0 Posts
  • 39 Comments
Joined 5 months ago
cake
Cake day: February 14th, 2026

help-circle





  • From what I’ve noticed, yes. Considerably.

    I’m not knowledgeable enough to explain why, but something about running Baremetal --> VM --> Docker --> Nextcloud-AIO is massively slower than running Baremetal --> Docker --> Nextcloud-AIO. Hell, Nextcloud-AIO on a Pi4 was running faster than when I put it in a much roomier VM.

    Someone tried to explain it to me but all I understood was that the databases don’t like that. Something about nested virtualization restricting performance.

    Oddly I didn’t run into the same issue when I ran Nextcloud-AIO off of a Digital Ocean VPS. Not sure what they are doing differently, but that was running just as fast as bare metal.









  • With the caveat that I last played with Gentoo 20 years ago… I am almost certainly a bit out of date.

    If I remember correctly it, it explicitly recommended that you use at least the minimal gentoo live disk to get your system into a running state. You’d be working from the live cd for the first couple of sections before booting into a very basic install on your hard disk. From there you would compile the rest of your system.

    Even the minimal disk provides all of the tools that you need to bootstrap the system. Sources for everything else are downloaded as they are needed. Come to think of it, I think the full desktop live dvd was fairly new at that time, in it’s first or second release.

    Even at that time the Gentoo manual was incredibly well written and is in my opinion the gold standard for how user documentation should be written. I had been toying with linux for about 3 months at that point and was able to get a working desktop system up and running in about a month , mostly just waiting for things to compile on the slow processors we had back then. I would run a few commands and then go off and do something else for a few hours. rinse and repeat.


  • Linux has run on ARM procs for some time. Software is a little hit or miss, but most things have a compiled build for it at this point. A lot of the big servers are running ARM processors due to potential power savings.

    The popularity of the Raspberry Pi really increased the number of projects with ARM builds as well. It’s been possible to run a pretty decent desktop stack for 10 or 15 years. When the Pi2B came out.

    If you happen to run across a project that is not available on ARM you might give a go at compiling it yourself. About half of the time it’s not too difficult and a good beginner project.





  • Meh. Reminds me a bit of the kerflufel when SystemD came out and largely replaced the V init system that came before. A whole bunch of religious adjacent arguments, at high volume with not much intelligence or understanding. It’ll pass.

    All I need to know is does it solve a problem I have, does it work, is it stable, and is it secure.

    Only warning I’ll give is that you should probably not get too used to your off site LLM models (Claude, GPT or whatever you’re using). Pricing seems unsustainable and the hype makes it feel like a bubble similar to the dot com bubble.

    Might want to devote some time to trying to bring your LLM usage in-house. There is no telling who will survive the crash and it’s not always the “best” one.