• 0 Posts
  • 19 Comments
Joined 3 years ago
cake
Cake day: July 29th, 2023

help-circle

  • vLLM to host and openwebui for a frontend that I really only use to demo models to friends.

    harness

    Opencode running with full access inside a rootless podman container with only the directories I want mounted (usually a single project, sometimes read only).

    So far though, I’ve only used it to document and report on project repos because Qwen is not that good at actually coding (although it can give pretty decent conf file examples, but not always optimal).

    Also useful for parsing pita data dump type files that I’m too lazy to construct multiline awk commands for.



  • Uh no, most if not all of the paid indexers are not worth it compared to just waiting for an open invite from a quality one.

    Not even the gold tier quality ones allow paid entry AFAIK. They only allow people who are long time seeders and uploaders with reputation.

    The one thing you might have to pay for is a VPN in case you want to torrent on public indexers, which is often worth it because not everything will end up on private indexers, and you usually have to wait a couple of months before you can join a solid private indexer.


  • I can trip the measly 15 amp breaker by drawing 1.5kW on two servers + turning on an amplifier that is somehow on the same circuit despite being in a different room quite far from where my servers are lol.

    On average though I only draw like 350W. The huge spike only happens if I run a miner/llm on both of my servers or the during the startup spike.

    The 1kW chunk comes from some RTX Quadro 6000s which each eat 250W each when maxxed out.

    EDIT: Want to mention that most machines report their power draw (from mobo) without IPMI, so you can use your favorite stats collector use it to form a rough estimate of kWh total in case you don’t have a proper meter that records usage.



  • I actually just did this recently with a BTRFS subvolume using a swap file and LUKS encryption.

    The old Fedora magazine article has the overall steps required and someone also made a guide for secure boot on I think the spiceworks framework forum.

    The catch is that secure boot will only allow signed images to be loaded which includes the hibernation swap image, meaning you have to use a customized kernel that allows unsigned hibernation images to load.

    I’ll share both links when I get a chance.

    EDIT:

    Fedora Magazine guide: https://fedoramagazine.org/hibernation-in-fedora-36-workstation/

    This is slightly outdated as Fedora now uses a proper zram.swap definition with systemd, so you can also do a systemctl edit to disable and enable zram0 with post/pre start on the hibernate service directly instead of using swapon/swapoff. I’ll make a second edit with my example I’ll have to dig out of some poorly documented notes lol.

    https://community.frame.work/t/guide-fedora-36-hibernation-with-enabled-secure-boot-and-full-disk-encryption-fde-decrypting-over-tpm2/25474

    This is the full forum post guide for FDE + TPM + secure boot. I’m not entirely sure of the kernel patch specific but it seems that the hibernate swap image is technicallly signed because if the bootloader process, but the kernel lockdown mode rejects it regardless by default because there’s no guarantee that the swap image was configured with FDE and not tampered with, which would be a vulnerability.

    Quick glance at the kernel patch makes it seem like it’s just a flag check to allow hibernate in lockdown mode, so I assume the secure boot signature check is still completed.

    lockdown_hibernate [HIBERNATION] Enable hibernation even if lockdown is enabled. Enable this only if your swap is encrypted and secured properly, as an attacker can modify the kernel offline during hibernation.


  • He actually offered to return to the US if they accepted that he would be given a public trial and not a military trial.

    Obama’s administration rejected the idea and said that their only guarantee they would give was that they wouldn’t torture him.

    Just to add my 2 cents: the list of countries that the US can’t extradite is not that long, especially when you exposed government corruption of the highest degree and you are considered a traitor.

    EDIT:

    Just to add my 2 cents: the list of countries that the US can’t extradite is not that long, especially when you exposed government corruption of the highest degree and you are considered a traitor.

    Just to add, Snowden also indirectly exposed the US’s leverage in several countries that you would otherwise assume would not cooperate with the US in any capacity.

    The fact that he ended up in Russia was because he couldn’t proceed to Ecuador which was his original planned location of asylum after leaving to Hong Kong.

    There are really very few countries the US can’t realistically coerce or run a black ops mission to go after him. China didn’t want the political baggage for no gain and Iran can’t even protect its own leaders from strategic bombing.

    That leaves basically Russia and its less than ideal former soviet allies.




  • I don’t want to run a separate database or anything complex, so things like Immich or Photoprism are not an option.

    Why not just use podman and let it handle itself? Unless you’re planning on hosting like maybe several million photos, I doubt you would ever need to manually set up a DB.

    Just download the compose file, add your photo directory paths to it, and let it start and never think about it again until you feel like updating for some new feature months later.




  • Saw the title and I already knew IPv6 was still gonna be “not enough adoption” lol.

    The overly aggressive lockdown from NAT is not entirely misplaced though. Windows is already notorious for default hosting an atrocious level of protocols that are anywhere between “I can get your NetNTLMv2 hash for free” to “gg ez RCE as SYSTEM user”.

    Linux and Unix were also not magically protected either, like every FTP server implementation has had several RCE vulns and it was so bad that Linus himself pointed out how easy it was to abuse buffer overflow vulns even after the addition of NX mem protection with return to libc, several years before ASLR finally landed in Linux and Windows.

    Seriously, the early 2000s was the wild west of vulns because everything was still written in C++ so you could just waltz into any binary, find some stack/heap overflow, and you’re on your way to an RCE.

    NAT (and firewall) most definitely did prevent millions of users from getting rekt by a metric ton of malware that operated on the ability to connect to those services.