It will be controlled by Truenas not Proxmox. Truenas can add swap space to each drive automatically: https://www.ixsystems.com/documentation/truenas/11.3-U2.2/storage.html

But you probably already have existing drives so that doesn’t help. This might though: https://wiki.debian.org/Swap

But be aware that Truenas is design to be an appliance and doesn’t really want you tinkering under the hood. So you may have to manually add the SWAP after each boot of TN.

I would guess the best long term fix would be moving services out of the TN VM and into a different VM.

Maybe start by taking an existing script you wrote in another language and hand rewrite it in C? Then you can focus on understanding how things are done differently in C.

I have TN Scale VM hosted in Proxmox. The only “issue” I have is the webgui gets pushed to SWAP if not used for more than a week. So when I connect it it literally takes a couple minutes while is gets shuffled back into RAM. Once it’s “warmed up” it’s fine. But my Scale VM is doing these things: manage ZFS pools, control NFS/Samba shares, replicate pool snapshots to off-site backup server. It intentionally have it do nothing else. All other services are in different VMs or LXC containers in Proxmox.

Does your Scale install have any SWAP space setup? That should prevent out of memory issues. Potential performance issues would be better than crashing.

I can understand the potential problems of trying to define “low-effort post”. In contrast can guidelines be given for a “quality post”. If no guidance on either end is given it may discourage some people from posting anything. Maybe people can contribute what they see as indications of a “quality post”.

I don’t see how it would push manufacturers to do that. I can see how it would make consumers more open to soldered RAM if RAM is so expensive there is no way you are going to upgrade it later. But, I would be interested to get your thoughts as I miss stuff that feels obvious I’m hindsight all the time.

Yeah. Basically just acting as a pretty good search engine to find relevant quotes.

Brandon Sanderson talked about it in the context art, “Making art with AI doesn’t make you an artist, it makes you an art director.” People who only understand grunt level coding work are at risk. The work in the future will require higher level skill sets in engineering, architecture, and project management.

Just link them the story of a Dad getting locked out of his Google account after sending a picture of his child to the Doctor as part of remote care.

https://www.theguardian.com/technology/2022/aug/22/google-csam-account-blocked

The point being they can fuck up your life on a whim and don’t care about the harm they will cause because your one out of millions.

My company has internally hosted AI. I use the web interface to copy/past info between it and my IDE. So far I have gotten best results from uploading the official Python documentation and the documentation for the framework I am using. I then specify my requirements, review the output, and either use the code or request a new revision with information on what I want it to correct. I generally focus on requesting smaller focused bits of code. Though that may be for my benefit so I can make sure I understand what everything is doing.

This is an area where AI can be helpful. Tell the AI what Linux distro your on and what you want to do. Most of the time it will give you pretty good answer. If you don’t understand what it is telling you to do, ask it to explain the thing in detail. Most important thing though is to always verify what it tells you before you run stuff. Google search specific commands or use the “man” command to get documentation. The key thing is the AI can make you aware of CLI commands and tools more easily then trying to find what you need on your own.

I think you are making it more complicated than needed. If you just want reliable service, just figure out who has the most reliable Internet and power and they can host the server. If you want to learn kubernets or docker swarm, you can try that but it will take a ton of upfront work.

Edit: Also get a UPS for the server.

Be aware if you have the iGPU as the only video output device and passthrough to a VM it will no longer show what the host system is doing. This would be referred to as a headless server. I would suggest making sure you can SSH into the host before doing that. LearnLinuxTV has guides for how to do that with best security practices.

Docker works differently, so it may not be an issue with that.

Truenas uses KVM for virtual machines. So that will allow GPU passthrough, but may require command line and config files to do it. For docker this seems relevant: https://forums.truenas.com/t/electric-eel-nvidia-gpu-passthrough-support/11797

When you read files from the ZFS filesystem it will automatically keep the files in RAM. This is called the ARC and it is why people frequently recommend having a lot of RAM with ZFS. The ARC is very effective, automatic, and has no risk because it only caches reads. A cache drive is a secondary ARC generally using a fast SSD. The problem is that it generally only helps performance when you are reading lots of small files multiple times. This is because ZFS does so well reading large files from HDD that it doesn’t make much of a difference.

In short: If you already have the drive and want to play with the feature, go for it. But if your going to spend money on the drive, you will probably be better served spending it on more RAM.

I assume the cache drive is for your ZFS pool. You probably don’t need that.

What about GiB?

Intel iGPU are very good for transcoding.

Other people have already talked about why you are having performance issues with the Pi. As for a better NAS solution you will probably be better off with a used desktop PC from the last 10 years. If the computer doesn’t have enough SATA ports you can get a sata addon card or HBA (host buss adapter) addon card flashed in IT mode. You should be able to find a lot of options on eBay. Maybe people can chime in with specific models to look at.

That sounds like a problem for the people dumping money into these companies and keeping them afloat.

I feel like this is a normal cycle of new tech. People get really excited about all the possibilities and don’t have any experience to ground expectations. Eventually people use it enough to realize what is more realistically achievable. Then mentally shifts from “magical solution to everything” to “a tool that is good at some things and bad at others”.