Does it even fucking matter what’s banned in what echo chamber??

Thanks. My setup is way over complicated with 3 hosts in a cluster and shared storage, so local storage on the hosts stay unused. But i have been thinking about redoing it with separate hosts. This solution looks promising for sharing data, even if just on one host

So lxc containers and not vm’s

Could you explain further with a bit more detail? I havnt looked at this in a while but back then the options where virtiofs or nfs

If you want to share storage you need some way of doing that. Zfs is a good option for storage on vm-host level, but ist not designed for shared usage. Im not sure what you are after, but maybe you want zfs storage inside the vm for snapshots, dedup etc? Or maybe you want to share your media storage between vm’s? The first case you can use zfs inside your vm, it does not know or care about how its disks are stored or of they are a physical drive. For the second use-case you want some way to share drives, like smb, nfs etc. or a distributed filesystem if you really want to over complicate things. Truenas might be over overkill for sharing a few volumes, but you need something. I believe you can share zfs over nfs now but i have never used that outside of proxmox cluster storage

I use both debian on a vm with samba+nfs and a bare metal truenas for my needs. Find your needs and figure out what solves them

So you mount the pool to each vm that needs the shared data? Afaik zfs is not made for concurrency

Looks like a tiny screw hole on the top.

Scrap that - after upgrading it went bonkers and will always use one of my «knowledges» no matter what I try. The websearch fails even with ddg as engine. Its aways seemed like the ui was made by unskilled labour, but this is just horrible. 2/10 not recommended

Possibly. Been running it since last summer, but like i say the small models dont do much good for me. I have tried llama3.1 olmo2, deepseek r1 in a few variants, qwen2. Qwen2.5 coder, mistral, codellama, starcoder2, nemotron-mini, llama3.2, qwen2.5-coder, gamma2 and llava.

I use perplexity and mistral as paid, with much better quality. Openwebui is great though, but my hardware is lacking

Edit: saw that my mate is still using it a bit so i’ll update openwebu frpm 0.4 to 0.5.20 for him. Hes a bit anxious about sending data to the cloud so he dont mind the quality

I have the same setup, but its not very usable as my graphics card has 6gb ram. I want one with 20 or 24, as the 6b models are pain and the tiny ones don’t give me much.

Ollama was pretty easy to set up on windows, and its eqsy to download and test the models ollama has available

No - i would advocate for not using docker if I need a network interface. But thats my opinion, and others will have a different one.

You can use macvlan networking, and if you need host<->container communication you give your host a macvlan interface instead or in addition to the root nic. Macvlan works “on top of” an existing interface, so theres no routing locally between the underlying nic and the macvlan nics.

If the host have several nic’s you can pass one through to a given container

There are other solutions than docker for that use-case that I think are better fits. It probably works fine, but for me other drivers including host mode and ipvlan seems to have been introduced to solve the wrong thing. Like how it needs privilege for them to work and how it exposes the containers network interface. For me it kinda breaks parts of why i would use docker.

Its my personal opinion and how i like to work.

You could probably make your setup work but it seems too complicated for me when you introduce a bridge as the root interface. Maybe with macvlan adapters on the host instead or in addition.

I dont get it - are you trying to mimic vm’s with you docker containers? docker works great using the normal way of exposing ports from the internal docker net through the host. Making technology work in ways it wasnt designed for usually gives you a hard to maintain setup

You have a mighty big hand if you reach l and a with the same one

I must have been doing websites wrong for decades by not forwarding through a reverse proxy. I admit a good one like caddy makes tls easy, but unless you have several backends for one site theres no need. The reverse proxy part of load balancer is very similar to a port forward, but on layer 7 instead of 3 or 4

There are other things to consider as well. Nfsv3 is good for large sequential reads/writes. There are no multichannel in nfs3 and no caching , and you should adapt technology to the use-case. For vm storage nfs sucks while for movie storage it works great.

For general file storage I would pick smbv3 for speed and ease of user/security

Multicast DNS uses multicasting (surprise!) so keep your devices on the same network and it just works. Docker is not very multicast friendly but lxc or a vm should have no issues.

By using its ip:port

If you have homeserver1 running stuff and that’s connected to the vps through vpn you make sure your homeserver1 service that you want to proxy runs on the homeserver1’s vpn ip.

For docker this is done by specifying that ip when you expose ports or use 0.0.0.0

I assume you can already ping homeserver1 from the vps by using the vpn address of homeserver1

I’d say it depends. Some patterns will be different. Like trying to create zig code when the LLM is trained mostly on javascript and python.

You dont need or want docker on your vm host. But a bare metal docker host can solve many peoples needs.