oscardejarjayes [comrade/them]

  • 0 Posts
  • 88 Comments
Joined 2 years ago
cake
Cake day: September 21st, 2023

help-circle
  • Wasn’t exactly my workplace, but a contractor. Basically, as a cost saving measure, they layed off half of the IT department. And then they got hacked. They just re-flashed everything, and the threat was out of their system, but they messed up big time. The new images weren’t locked down properly, so they almost immediately got hacked again. I noticed that they’d messed up, and pointed it out to a few people, but it was too late.

    Now the execs need a scapegoat, so they gut the IT department again. I don’t work for them, not even close to the business relationship, but their managers call me to a meeting room and try to get me in trouble? Try to make me admit to doing something wrong? And it was just their admin people there, not like my heads or anything. It was kind of a surreal experience.

    This was a while ago, and their tech is still a bit funky. (Some details are lightly fuzzed, but this all is basically true)




  • restic to a local server and to cloud storage. it varies by device, but usually just everything in /home/. The rest of the operating system should be reproducible, whether through images, ansible, nix, or guix, given the information in /home/.

    scheduling is done through systemd, usually (or the non-systemd equivalent). I use BackBlaze now, but I switch around occasionally. restic has policy based snapshot removal, and a prune option.



  • You could try using Hashicorp’s Packer to generate images repeatably (usually more meant for cloud images though). Or NixOS (like others have mention), or Guix (like NixOS, but better in some ways, worse in others). You could make it an Ansible playbook, which would let you both make configured images, and just configure machines that already have an OS.

    I do something similar with archiso, fwiw, but that only works with Arch Linux.

    Would you want to change your distribution, or just keep Debian with some tools to automate?




  • It’s not really something we can do, sadly. Reddit closing it’s API was more about getting money than actually stopping it’s use as a training set.

    Having an allow-list is a start though, as it means that a company can’t just make an instance and suck all the data out through that. Common corporate crawlers could be added to the robots.txt, but that would mean that you might not be able to find lemmy instances in search results. We could make it against ToS, but what are we going to do, sue the massive corporation? They have plenty of lawyer and payout money, so very little would fundamentally change.

    Ultimately, if content can be served to us, it can be served to them.


  • Exciting! Sort of interestingly, I never dual booted or anything, I just jumped straight to Linux.

    Honestly, it’s really not that bad. Linux has come a long way since I started out, and while I usually make it harder for myself than it needs to be, I’ve seen young middle schoolers installing and using Linux, I’ve seen retired professional musicians with no technical background install and use Linux. Especially with all these new fancy atomic desktops, like Silverblue, Bazzite, and Kinoite. Admittedly, I have managed to break a Kinoite installation (doing stuff I probably shouldn’t have been doing), but fixing it felt magical. Just roll back to when it wasn’t borked, then update it.

    I did a lot of not so nice things to that installation (it was a bit of a test, to see how fragile it was), and it’s still running now!


  • I almost thought you were that bot that changes youtube links to invidious ones, lol.

    Yeah, those tend to be good (well, tux.pizza is a bit of an exception, it shows the error that the others fixed). It’s a little annoying that a lot of the invidious instances that work won’t show up when you do the “switch instance” thing on an instance that doesn’t work, but it makes a bit of sense, not wanting to get overwhelmed, or trying to not get too noticed.