It’s been a while! Let us know what you’re currently working on, what exploded, what cool stuff you found, or just brag that everything is running fine unattended.

I haven’t been doing a lot, renovate auto upgrades for patches and sometimes minor updates work great so far. I got involved in my union to spin up a mastodon server soon, so there’s a new project, but I already run mastodon so it shouldn’t be much of a deal, I just need to get comfortable with the infrastructure. I also set up ntfy but only sent some test notifications and didn’t configure any app yet - planned for Element Chat.

[Meta] I’m not sure how to apply the new tags. In the recent versions of this thread, people sometimes posted about their ollama setup or the like, but I’m not mentioning AI in the post body (except this meta thingy) and it shouldn’t be much of the discussion, but it’s an open post. Maybe curbstickle can chime in to let me know if I should apply one of them.

  • Victoria@lemmy.blahaj.zoneEnglish
    1·
    5 days ago

    Fiddling around with Renovate Bot on my Forgejo instance. It’s not working perfectly yet, but it’s getting there.

    Also I began using LubeLogger to track my car expenses and fuel efficiency. Works beautifully.

      • pienoyer@piefed.socialEnglish
        3·
        4 days ago

        Can only speak for myself. I went with PocketID for oidc. The learning curve was a little steeper than most other homelab services i try, but its 111% worth the setup and time. Logging into services now happens automatically. And if you’re distributing services to friends and family then oidc is a lot more seamless than 10 different accounts on 10 different services.

        I went with PocketID because it looked the simplest out of the bunch. Authentik does provide more customization but afaik this is more needed for enterprise graded stuff.

        All in all, very happy with it, not going back to traditional logins any time soon.

        • Alfredolin@sopuli.xyzEnglish
          2·
          3 days ago

          Woah, thanks for the tip.

          I just tried pocketid and it’s actually not that complicated. Nice.

          • Alfredolin@sopuli.xyzEnglish
            1·
            2 days ago

            I mean spooling up pocket id was simple. Putting oidc everywhere IS fastidious but awesome.

      • Drunk & Root@sh.itjust.worksEnglish
        1·
        4 days ago

        It was a pretty huge learning curve but it was great learning to troubleshoot stuff i also set up oauth2-proxy and that was also a bit of learning curve but worth it in the end

  • kokomo@reddit.kokomo.cloudEnglish
    1·
    5 days ago

    Not to brag but everything’s working fine, clean as a whistle. Upgraded forgejo to the latest version. I’m still helping maintain the awesome-selfhosted list, and get inspiration from selfh.st a lot (big shoutout!) Hope to make sure some postgres dbs are up to date at some point, all of my dockhand instances that i set up have been a big help :)

  • Decronym@lemmy.decronym.xyzBEnglish
    1·
    2 days ago

    Acronyms, initialisms, abbreviations, contractions, and other phrases which expand to something larger, that I’ve seen in this thread:

    Fewer Letters More Letters
    AP WiFi Access Point
    DNS Domain Name Service/System
    Git Popular version control system, primarily for code
    ISP Internet Service Provider
    NAS Network-Attached Storage
    NVMe Non-Volatile Memory Express interface for mass storage
    PCIe Peripheral Component Interconnect Express
    PSU Power Supply Unit
    PoE Power over Ethernet
    SATA Serial AT Attachment interface for mass storage
    SSD Solid State Drive mass storage
    SSH Secure Shell for remote terminal access
    XMPP Extensible Messaging and Presence Protocol (‘Jabber’) for open instant messaging

    [Thread #72 for this comm, first seen 2nd Aug 2026, 22:30] [FAQ] [Full list] [Contact] [Source code]

  • SayCyberOnceMore@feddit.ukEnglish
    1·
    5 days ago

    I found the best thing to deal with a device that won’t work…

    I’m trying to get the 3.5" adafruit touchscreen working and it’s being stubborn…

    So… I went outside and split a large pile of logs.

    Problem still not fixed, but I feel a hellofalot better

  • brucethemoose@lemmy.worldEnglish
    1·
    5 days ago

    [AIT] I know this isn’t everyone’s cup of tea, but I’m excited about Deepseek V4 flash. It’s (for me) the perfect size and architecture to self-host an LLM.

    My box (and brain) are chugging through a queue:

    • Figure out why my swap is going crazy, and how to ban processes from it [Done].

    • Figure out why Code OSS is unhappy [Partially Done].

    • Make an ik_llama.cpp iMatrix for Deepseek V4 [Done].

    • Figure out why quantization isn’t working [Done].

    • Make a test IQ2_KL/MXFP4_R8 quant to see how it does squeezed onto my box [in progress].

    • Test. Tune. Inevitably troubleshoot the dozen other things that go wrong. Figure out how much spare RAM that leaves me.

    • Make a higher quality IQ3_KT quantization. This will take all night on my CPU.

    • KLD test both of them vs the full precision, to quantify quantization loss. Likely an overnight task, too.

    • Try merging the new model release with the base model, 50/50, for a less “deep fried” model. imatrix, quant, test.

    The goal is to host it on a single RTX 3090, Ryzen 7000 with 128GB RAM, for anyone curious. Though I may try smaller models too, like Laguna S1.

      • brucethemoose@lemmy.worldEnglish
        1·
        5 days ago

        • It’s just under 300B, trained at FP4; absolutely the perfect size for servers with 128GB-192GB CPU RAM to spare.

        • Its fast. I’m getting 17 tokens/sec on a single RTX 3090 GPU, all experts offloaded to RAM; for a 300B model this smart, that’s crazy fast.

        • Its attention mechanism is cutting edge, good for long context without too much processing time.

        • The benchmarks for coding/agenic usage are absolutely bonkers, within margin of error of frontier models or Deepseek Pro in some cases: https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731

        • …Though I don’t put much stock in benchmarks. And I haven’t tested it enough to tell you if it lives up to that hype for specific use cases.

        • Deepseek also publishes its base model. That means I can “unfry” the model with a merge if I have to.

        I am afraid of the the model being “overfit” to coding and agenic stuff.

        For reference, my previous favorite model was Xiaomi MiMo V2.5 310B. It benched well, but it also feels “smart” outside of benchmarks, like in knowledge of trivia without tool usage/internet access or comprehension of weird questions.

  • Rin@lemmy.zipEnglish
    0·
    5 days ago

    Finally think I got everything installed on my Raspberry Pi 4 that I want there, thankfully.

    I’ve been trying to document my journey with selfhosting this time better than previous attempts, as far as writing down what worked and didn’t work. It’s coupled with me trying to write little explanation guides for my partner about what service does what, and how to use them. They’re really excited to actually use them, but I want to get the guides done first, which has been a challenge.

    Kinda surprised Grimmory is taking up so much RAM but if it becomes an issue I’ll put it on my other, beefier server. It was the last service I was struggling with, because every time I would upload a book, something would go wrong. An example being that one time I uploaded a book and it went into every single folder after, then deleting one copy deleted all the copies.

    Admittedly, I’m not terribly well experienced so it was probably an issue on my end. Still learning!

    Speaking of, my next project to tackle is overhauling my more powerful server and what it’s running. Since I use a split setup where I store media on my NAS, I’ve had difficulties with databases hating that aspect. Now that I know a bit more than I did when I started, and I have nowhere else to ask (thank you for having this community, by the way, it seems a lot more newbie friendly than some others) I was wondering if I could map the database of a service to be on the server itself, while having the media mapped to my NAS? From my research, I think this is possible, but I’m not 100% sure.

    Hope everyone else’s experiences are going well!

    • glizzyguzzler@piefed.blahaj.zoneEnglish
      0·
      4 days ago

      Check out BookOrbit if you find Grimmory too hefty in the RAM reqs. It uses very little RAM for me. I think it also has import from Grimmory as well.

      As for the database Q, if you’re using docker/Podman it’ll be very easy to point the database folder (if included in the image, like Jellyfin does) to your local filesystem and the media to your NAS. Same idea if the container uses a Postgres container, just spread across two containers in that case.

      • Rin@lemmy.zipEnglish
        0·
        4 days ago

        First, thank you much for suggesting BookOrbit! I actually just installed it because I was having a lot of issues with Grimmory making dupes of books in every folder and couldn’t figure out why it was doing it. Haven’t had any issues with BookOrbit and I really like that it’s got achievements. Plus, the less RAM usage is nice.

        I’m using docker and docker compose files! Thank you for answering, the last time I took a crack at selfhosting, I really felt lost and screwed a lot up, so I wasn’t sure if that was even possible.

        • glizzyguzzler@piefed.blahaj.zoneEnglish
          1·
          4 days ago

          The docker compose file is great, clear once you get used to all its little sections. For volumes I only use what they call “bind mounts” which are where you have a folder on your system connected to the folder in the container. As opposed to the “docker volumes” which are internal to docker.

          And with those it’s super easy to just be like in the volume subsection:

          - /path/to/local/drive:/container/database:rw,noexec,nosuid,nodev,Z

          - /path/to/network/drive:/container/media:rw,noexec,nosuid,nodev,Z

          The :rw,noexec,nosuid,nodev,Z at the end is a great extra security thing that never causes problems. rw means read-write, you can switch it to ro for read-only if you’ve got something you want the container to only read from but not be able to modify. I use that for jellyfin’s media since I don’t want it doing anything but reading it. The noexec means don’t let executables be run from the folder, never should happen so it just prevents a sick hack from being put in the folder and run. I forget what the others do but they’ve never been a bother. And Z means only one process can access the volume, you can switch it to lower case z to let multiple processes access the volume - and I’m not sure it does anything without SELinux going which I think only fedora does by default right now.

          Enjoy the info dump!

  • neidu3@sh.itjust.worksEnglish
    0·
    5 days ago

    Not working on anything particular, but I’m going on a road trip on Tuesday to this regional airport that is giving away two racks perfect for my use case. Deep enough for my larger machines, sound proofed and 42U tall.

    Finally my hardware can have a proper rack instead of being in a stack/pile in this harrypotteresque locker under my stairs.