I think I get it. I mean in that situation you’d essentially pay to get some SATA ports and the space to put the harddrives. The money doesn’t really get you anything else that’d be fundamentally different from the current setup.

Idk, I’m fine with 48GB of RAM to run a lot of services and containers. And I don’t use a separate machine for storage, the hypervisor does that and I either share the filesystems via NFS or pass them through into some VM. And I don’t think a fast machine with lots of RAM is needed for storage, unless you’re using ZFS.

The HDDs alone should be roughly ~1,000€. The rest of the build sounds pretty much like your other machine, just with a different processor.

I run my YunoHost in a VM with like ~8GB of RAM allocated. You can move everything to one single machine if you set up some reverse proxy for all the web frontends.

36TB of storage and 64GB of RAM should be plenty.

I’d say it’s good if it’s easy to use, well written with maintainability in mind, offers good functionality, is reliable and follows current best practices.

It’s easy to selfhost if it’s packaged. Because then I can just apt install gitlab edit a few config files and I’m done. Or click on it in Yunohost, or maybe run the Docker container.

But just “easy” isn’t the whole story. It needs to be maintainable, still around in a few years, integrate into the rest of my ecosystem…

I meant the other side. If you use 200GB of your Proxmox, you don’t need to transfer it to the NAS. Which is the question here. I don’t do it, because it’s mostly calendars, contacts and like 10GB of data on Nextcloud, which I’m currently working on, or sharing with friends. And the “NAS” is sleeping most of the day. But if OP wants all their data stored on a NAS, they might very well configure Nextcloud to use NFS and do it that way.

Sure. That’d be a valid use-case. I don’t think I can recommended anything, here. Both should work fine. And you can always run into some unforeseen consequences in a few years. Especially once you decide to change something about your setup. But these things are hard to factor in. I often tend to prefer the easier solution over the more complex one. That helps with maintainability. But that approach doesn’t always apply.

Edit: If you want everything stored on the NAS, just do it.

https://lemmyverse.net/communities

Or copying subscriptions from another (old) account someplace else.

Is the storage shared with other software, or are the NAS and Proxmox two different machines? Or why bother to set up NFS at all?

I keep most of my usual data local. But I’ve also added some external storage to Nextcloud, that’d be a large harddisk that contains some music and TV recordings, Linux ISOs and rarely used stuff. It’s mounted read-only most of the times and spins down unless I need to access my achived stuff. That’s the main reason why I keep it separate.

True. Also the whole atmosphere of the community matters. And how useful a platform is. And not everybody likes arguing about things. But passively consuming videos is a different story. And the videos are embedded into other websites and Lemmy isn’t. Also a that size every single creator and their followers make a difference.

I got a Xiaomi scale 2 (years ago) and that works with the OpenScale app. It’s Free Software. There is a list of supported bluetooth low-energy scales by that project/app.

I’d like to know what people use Lemmy for… Reading the news, scrolling through meme pictures or talking about hobbies…

If they subscribe to communities tailored to their interests or just watch the “All” feed.

And maybe moderation. If users have previous experience with moderation and whether they think their experience with that was negative or positive.

And I’m always interested to know if people still use computers. And maybe the operating system, to judge the amount of Linux nerds here.

I’m not seeing that trend. There are people who don’t want to federate. But they usually choose different software. Like Flarum or Discourse instead of Lemmy. And I think the statistics pages pull the user numbers directly from the instances(?) so a more defederated instance might show up in the statistics anyways.

Agree. The idea of having a large, joined world is nice. Like having some portals or borders which lead to the next server… But it changes things and is difficult to implement and integrate into the world and gameplay mechanics. On the other hand Minetest/Luanti is great for tinkering. So it would fit there. It’s just complicated. And we don’t really “need” it. And the light version: just federating usernames or inventory doesn’t do much for the gameplay.

Nice. We’ve had that idea in Luanti, but it didn’t get accepted /merged.

It should automatically add a subscription and show that, after Element picked up on it. Otherwise it won’t do anything. I don’t remember if I had to do any additional steps.

Uh, I’m not sure about that setup. Usually F-Droid strip the Google libraries, so you won’t get push notifications via whatever the default Google mechanism is. (And I’m not sure if it also requres some setup on the server side.)

And without push notifications, you’ll experience severely delayed notifications. I installed ntfy for that. That’s an alternative push notification provider. And that works for the F-Droid versions of FluffyChat and Element X and such…

Maybe try the Play-Store version of Element and see if that works. Or install something like ntfy on your phone if you like the F-Droid builds better.

I don’t know much about Dendrite, but maybe you want to add which client on what kind of device you’re using… Computer or phone is kind of a different story concerning notifications. And on Android it also matters which brand and if you’re using an aftermarket OS variant maybe without the Google services… or the stock Android that came with the phone. I’m just asking since we’re in a nerd community here…

A VM is practically as secure as a dedicated machine. I mean in theory it isn’t. But in practice, that’s how everybody does it, including the big tech companies. And there’s rarely any dangerous vulnerabilities.

It all depends on how you set it up. If the machines lack proper firewalling and can reach internal services. If you didn’t set some permissions right. Or there is a vulnerability in the software or the way it’s installed… I think that’d be the main concerns. And it doesn’t matter too much which exact virtualization or containerization setup you choose, as long as it provides isolation, and also isolation from the networks it isn’t supposed to access.

Just make sure you actually enable the jails/filters for the services you use … I’ve seen people just install it and that will by default just protect ssh and leave everything else as is.

That looks like the internet. Every server gets bombarded with these requests. Generally: Use good passwords, make sure your software blocks bots brute-forcing passwords, after some sane amount of tries… Keep everything updated…

If you want some more attacks, install a mailserver. Or expose VNC/Windows Remote Desktop or a VOIP server. That gets the bots really worked up.

Yeah, It’s either 4 lines and you got some service running… Or you need to learn a functional language, fight the software project and make it behave on an immutable filesystem and google 2 pages of boilerplate code to package it… I rarely had anything in-between. 😆