Build anything small into a container on your laptop, push it to DockerHub or the Github package registry then host it on fly.io for free.

Great write up, thanks. For video learners, Wolfgang does a good step-by-step on YouTube

I’d love you to check back later with your conclusions.

Guide to Self Hosting LLMs with Ollama.

• Download and run Ollama
• Open a terminal, type ollama run llama3.2

If it’s an M1, you def can and it will work great. With Ollama.

Thanks, I ended up going with Garage, but it has the same issue. I assumed I could just specify some buckets with their keys in the docker-compose or garage.toml, but no - they had to be done through the api or command line.

This is correct, I already installed the minio cli, but when I came back and read this, I tried it out and yes, once garage is running in the container, you can

alias garage="docker exec -ti <container name> /garage"

so you can do the cli things like garage bucket info test-bucket or whatever. The --help for the garage command is pretty great, which is good since they don’t write it up much in the docs.

Thanks. I ended up going with Garage (in Docker), and installed the minio client cli for these tasks.

One I’m writing. I use the host file system (as I have a strong preference for simple) for it’s storage, but I’m interested in adding Litestream for replicating the database onto AWS.

Love the effort you’ve put into this question. You’ve clearly done some quality research and thinking.

When I asked myself this same question a couple of years ago, I ended up just buying a second hand Synology NAS to use alongside my mini-pc. That would meet your criteria, and avoids the (I’m not sure what magnitude) reliability risk of using disks connected over USB. It’s more proprietary than I’d like, but it’s battle tested and reliable for me.

starcoder2:latest       	f67ae0f64584	1.7 GB	3 days ago 	
phi3:latest             	d184c916657e	2.2 GB	3 weeks ago	
deepseek-coder-v2:latest	8577f96d693e	8.9 GB	3 weeks ago	
llama3:8b-instruct-q8_0 	1b8e49cece7f	8.5 GB	3 weeks ago	
dolphin-mistral:latest  	5dc8c5a2be65	4.1 GB	3 weeks ago	
codeqwen:latest         	df352abf55b1	4.2 GB	3 weeks ago	
llama3:latest           	365c0bd3c000	4.7 GB	4 weeks ago

I mostly use starcoder2 with Continue for code autocomplete, the big deepseek coder is a bit slow (I can feel it thinking), but it and the regular llama3 are good for chatbot type programming questions.

I don’t really have anything to compare the M1 performance to. I guess the 8GB models output text a little slower than the web versions of the same models, and the 4GB ones about the same. Using ollama in the terminal, there’s sometimes a 0.5-2 second pause before it starts outputting. Not with phi3 though - it’s surprisingly snappy for the quality of answers.

An M1 MacBook with 16GB cheerfully runs llama3:8b outputting about 5 words a second. A second hand MacBook like that probably costs half to a third of a secondhand RTX3090.

It must suck to be a bargain hunting gamer. First bitcoin, and now AI.

edit: a letter

I use the Continue VS Code plugin with Ollama to use a couple of different models (deepseek-coder-v2 & starcoder2) to recreate a local only Github Copilot type experience for coding. This is on an M1 Apple Silicon though. For autocomplete the generation needs to be pretty brisk - I’m not sure how that would go in a VM without a GPU.

Yep, I think there’s sound arguments for separating out your storage (NAS) and network (router/DNS/PiHole) infrastructure. After that, whatever suits your purpose. I virtualise all my serious services on one machine under Proxmox (mostly for ease of snapshots) then have another machine for things I’m fiddling with, usually again under Proxmox so they are easy to move to production when I’m happy with them.

My NAS and production server run 24/7, I’ve got a dev server that I turn off if I’m not expecting to use it for a week or so. Usually when I do that, I immediately need it for something and I’m away from home. I have chosen equipment to try and minimize energy use to allow for constant running.

My view on UPS is it’s a crucial part of getting your availability percentage up. As my home lab turned into crucial services I used to replace commercial cloud options, that became more important to me. Whether it is to you will depend on what you’re running and why.

I’ve heard that one of the most likely times for hard drives to fail is on power up, and it also makes sense to me that the heating/cooling cycles would be bad for the magnetic coating, so my NAS is configured to keep them spinning, and it hasn’t been turned off since I last did a drive change.

I agree. Get a domain name, point it to the internal address of your NGINX Proxy manager (or other reverse proxy that manages certificates that you are used to). A bit of work initially, then trivial to add services afterwards.

I didn’t really need encryption for my internal services (although I guess that’s good), but I kept getting papercuts with browser warnings, not being able to save passwords, and some services (eg container repository on Forgejo) just flat out refusing to trust a http connection.

My step-up from Pi was to ebay HP 800 G1 minis then G2’s. They are really well made, there’s full repair manuals available, and they are just a pleasure to swap bits in and out. I’ve heard good things about, and expect similar build quality from the 1 liter Lenovos.

I agree that RAM is a likely constraint rather than processor for self-hosting workloads. Particularly in my case as I’m on Proxmox and run all my docker containers in separate LXCs. I run 32GB in the G2’s which was a straightforward upgrade (they take laptop like memory). One some of them I’ve upgraded the SSDs, or if not, I’ve added M.2 NVME drives (that the G2’s have a slot for).

Yeah na, put your home services in Tailscale, and for your VPS services set up the firewall for HTTP, HTTPS and SSH only, no root login, use keys, and run fail2ban to make hacking your SSH expensive. You’re a much smaller target than you think - really it’s just bots knocking on your door and they don’t have a profit motive for a DDOS.

From your description, I’d have the website on a VPS, and Immich at home behind TailScale. Job’s a goodun.

+1 for the main risk to my service reliability being me getting distracted by some other shiny thing and getting behind on maintenance.

I started as more “homelab” than “selfhosted” as first - so I was just stuffing around playing with things, but then that seemed sort of pointless and I wanted to run real workloads, then I discovered that was super useful and I loved extracting myself from commercial cloud services (dropbox etc). The point of this story is that I sort of built most of the infrastructure before I was running services that I (or family) depended on - which is where it can become a source of stress rather than fun, which is what I’m guessing you’re finding yourself in.

There’s no real way around this (the pressure you’re feeling), if you are running real services it is going to take some sysadmin work to get to the point where you feel relaxed that you can quickly deal with any problems. There’s lots of good advice elsewhere in this thread about bit and pieces to do this - the exact methods are going to vary according to your needs. Here’s mine (which is not perfect!).

• I’m running on a single mini PC & a Synology NAS setup for RAID 5
• I’ve got a nearly identical spare mini PC, and swap over to it for a couple of weeks (originally every month, but stretched out when I’m busy). That tests my ability to recover from that hardware failure.
• All my local workloads are in LXC containers or VM’s on Proxmox with automated snapshots that are my (bulky) backups, but allow for restoration in minutes if needed.
• The NAS is backed up locally to an external USB that’s not usually plugged in, and to a lower speced similar setup 300km away.
• All the workloads are dockerised, and I have a standard directory structure and compose approach so if I need to upgrade something or do some other maintenance of something I don’t often touch, I know where everything is with out looking back to the playbook
• I don’t use a script or Terrafrom to set those up, I’ve got a proxmox template with docker and tailscale etc installed that I use, so the only bit of unique infrastructure is the docker compose file which is source controlled on Forgejo
• Everything’s on UPSs
• A have a bunch of ansible playbooks for routine maintenance such as apt updates, also in source control
• all the VPS workloads are dockerised with the same directory structure, and behind NGINX PM. I’ve gotten super comfortable with one VPS provider, so that’s a weakness. I should try moving them one day. They are mostly static websites, plus one important web app that I have a tested backup strategy for, but not an automated one, so that needs addressed.
• I use a local and an external UptimeKuma for monitoring, enhanced by running a tiny server on every instance that just exposes a disk free and memory free api that can be consumed by Uptime.

I still have lots of single points of failure - Tailscale, my internet provider, my domain provider etc, but I think I’ve addressed the most common which would be hardware failures at home. My monitoring is also probably sub-par, I’m not really looking at logs unless I’m investigating a problem. Maybe there’s a Netdata or something in my future.

You’ve mentioned that a syncing to a remote server for backups is a step you don’t want to take, if you mean managing your own is a step you don’t want to take, then your solutions are a paid backup service like backblaze or, physically shuffling external USB drives (or extra NASs) back and forth to somewhere - depending on what downtime you can tolerate.