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.
You must log in or register to comment.
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
I built a dashboard themed around my dog. It uses the Home Assistant API to get various stats from the house. I’m pleased with it!
There’s not nearly enough dog metrics on that dashboard
Care to share any more details? Looks very aesthetically pleasing. I think I should make one themed around my cat.
It’s a PHP container On the front end I’m using Bootstrap and theming it with Catppuccin. The data for the home energy charts are populated from the Home Assistant API. They’re drawn with chart.js. The server stats are pulled from PHP exec commands, and drawn with Bootstrap progress bars. The temperature card also takes data from the Home Assistant API, which also has an endpoint for local weather data. This doesn’t update as often as I’d like. But it compares the outdoor temperature with the reported temperature of my thermostat, and then the messages below that are determined by a comparison of those values. I’m also using HTMX to load each card, and each tab of the energy card separately so that I don’t have no wait on all requests for the dashboard to load. Not that it’s slow - the home assistant API is surprisingly efficient at loading a great deal of logs. When you click my dogs face, she says WOOF WOOF WOOF in a modal to offer me a search with searxng.
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 [Done].
-
Make an ik_llama.cpp iMatrix for Deepseek V4 [Done].
-
Figure out what 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. 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.
-
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.
Deepseek V4 flash
What are the advantages of Deepseek V4 flash?
-
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.
Giving away? As in free? Luckyyy
Sound proofed… wow… does that also mean heat can’t escape?
Vent fans at the top and the bottom. But it can be next to my work desk without the hardware getting annoying.
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.
Renovate is awesome. Combined with Komodo you get a proper gitops style container management.
I set up oauth for my sites
How was the experience? And worth the pain?
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.
Woah, thanks for the tip.
I just tried pocketid and it’s actually not that complicated. Nice.
I finally got my immich docker running again. I lost all my meta data in the process but oh well, it works again.
what happened to make it stop running?
I think my Postgres docker became unsupported. I think Immich needs Postgres with vector to run and mine did not have vector. I tried to save my data but blew it up in the process. Nothing of great value was lost and I keep all my images elsewhere in my nas for just this reason. I do kinda wish Immich could just look at a directory rather than keeping its own file structure.
I thought I’d take a shortcut and use AI to help me setup an arch thing I’m working on. Jokes on me, took two days and in the end the thing was so fucked its irrecoverable. What a waste of a weekend.
What’s up? Not Grafana. I did
docker compose pullon the stack (without version pinning) then recreated the containers, and now it’s fucked. Or maybe it’s not because the logs are fine but it only serves the web app but 404s on any connection to the Grafana backend.Now I need to do some investigation into any recent changes to try to work out what’s going wrong.
Mealie, recipe manager and Social to Mealie. Scrapes recipes from instagram reels, tiktok, youtube shorts.
Bar Assistant, bar and cocktail management. You can add your ingredients and it shows what can you make, you can share a menu to your friends, its cool
Finally moved everything to the basement.
Spent years with everything in the office with me. Used to run some pretty old and noisy devices too lol. I always wanted to move everything somewhere else to cut down on noise but that means cables to run and total disassembly. Finally bit the bullet and moved it all last Saturday. The office has never been so peaceful
Building a new server, pinping up the specs to run more serious llm in parallel with all my services.
Waiting for the CPU (an amd Ryzen 9 5950X) that delayed in delivery, and building wooden crate/case to hold everything together with proper ventilation .
Ooo care to share pics of the case design? I’ve been interested in building one for awhile now
I have no design, but will surely make a post with a few pictures down the road, thank you for the interest!
It’s an open box approach, kind of a structure of wood around a metal motherboard plate, with place for disks and cable management, will keep taking pictures along the way and definitely make a post in a few days!
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 :)
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!
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.
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.
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,ZThe
:rw,noexec,nosuid,nodev,Zat 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!
Thank you again! I actually really appreciate the infodump and it was well explained. I’ve made notes on this in my selfhosting documentation so I can reference it later! I didn’t even know the difference between “docker volumes” and “bind mounts” really.
Setup an XMPP server for family and friends to use. While I really like it, I can’t seem to get Jitsi to play nice for group calls with the Openfire server. Atleast 1 on 1 calls seem to work fine with Dino and Conversations anyway.
