Good point about platform agnostic remote for management stuff. VNC is ideal for this.

And systems like Proxmox use a web GUI for most stuff, it’s a touch slow but I think that’s mostly just waiting for the system to finish the actual changes I make, and not the UI.

It uses some form of VNC (forget the name). Performance is fine for the VMs for non-video stuff.

You can run whatever you want inside a VM too.

Works for me, but damn it’s dog snot slow. Like typing takes 1-2 seconds for each character to show up

There was a post recently about “.LAN” recently being added to the DNS spec

Why would you run local traffic over the VPN?

I decided there simply were not enough docker apps for downloading Youtube videos, and so I made the situation worse :p

You get my upvote for this alone!

The other cost to home lab serving lots of data is internet limits.

My internet will throttle after X terabytes (I forget the limit), whether upstream or down, and of course upload is slower.

It’s like playing a game of tetris with costs, performance, stability, reliability, flexibility, privacy, security, and personal effort.

I’m not saying don’t have a home lab, just that there are things to consider. It’s worth the effort for me, though I’m working to move some things to cloud storage (e.g. Hetzner/Storj.io) and a VPS to get some of the bandwidth off my connection, and remove my home internet as a bottleneck or failure point.

That’s not much of an answer, I’m not reading docs because you can’t be bothered. I don’t use NixOS, so if you want to use that as an example, you’ll need to put in the effort to explain how it’s different.

If you don’t want to use LDAP, don’t. Then you get to manage each user account on each device.

To be frank, it seems like you have an adversarial attitude about this, and you think NixOS is the answer. Every one of your responses has been “but” whatever. You don’t seem like you want to understand how to use things, just complain it doesn’t work the way you think it should.

You need backup local admin accounts, not Backups for each user.

Which is how enterprise does things. There are local accounts with root access, but the id’s and passwords are tightly controlled.

Meaning what?

What’s wrong with LDAP for users? (I’m trying to think of a negative, and can’t).

I get downvoted to oblivion when I point out “just works” isn’t true.

You make a great point about endless choices.

No single UI, no single set of tools, those are massive barriers. And it’s why Windows became the de facto standard: single UI, consistent toolset.

“Rarely happens”

And yet it still happens often enough if you touch enough boxes or make enough changes across enough boxes.

The thing is, you never know when or where an unbootable box is going to occur. This is why I’ve carried a loaded thumbdrive for ~20 years (Lacie IAmAKey was my latest, wish they still made them - mine finally died. Knockoffs are available, unfortunately the casing is aluminum, not stainless, so they’re easy to bend). And why I keep hot spares around.

Host solution for what?

Each service you want will probably have a different set of options.

When you say Apple cloud, that could mean all sorts of things.

Specificity in tech is crucial.

You may want to start with one type of service, and go from there. You’re about to head down a deep rabbit hole that includes things like Security Posture, Risk Management, etc.

Funny someone downvoted you.

Clearly that person has never managed a 10,000 pc domain. Or hell, even a 10 pc domain in an SMB.

“The license is worth the cost” - I literally had this conversation with a peer not two hours ago. They have a client who’s previous IT management built a domain using Linux. Yes, you can do it, but I’d only do it if your IT is fully in-house and stable. This was an IT vendor. It saved them (the client) licensing…like $250 or so.

Imagine how quickly they’re going to burn $250 for a support issue because there’s something odd about how the Linux software isn’t exactly duplicating a windows DC? Or the next IT vendor doesn’t know what you implemented, so have to find out about which packages you used and how they work. (In this case they’re building a new domain and migrating everyone, because it’s currently unsupportable. Glad they saved $250 to spend $20k today).

You don’t use Linux desktop in a business to save licensing costs, unless you know the use-case inside and out. The first time your business has a need for something that doesn’t exist in Linux land, all those savings are gone as you build a virtual host for Windows, and deal with the lost productivity.

And I use Linux every day for things like Proxmox, UnRAID, TrueNAS, etc. Even there the difference between design approaches is really problematic.

This is the best description I’ve seen.

It gets old hearing the “Linux is better, Windows sucks” mantra.

They’re different things with different use cases.

I despise Linux for a desktop, it’s an awful experience, because it hasn’t been developed/targeted for what I need to do, and I don’t have the time to play fuck-fuck with distros to work something out - I have other shit to do.

Like build and manage Proxmox/TrueNAS boxes, which are… LINUX! Because this is where Linux shines, as purpose-built solutions.

Windows works nearly every time any more - I don’t have to do anything during setup. Drivers are automatic during setup.

Not sure where you get this idea from.

My Logitech mouse doesn’t work at all on Linux unless I search for why and go find third-party software for it. Windows sees it as a generic HID and treats it as such. I can go get the Logitech software if I want, but have no need of it. Linux? Nope. Probably the most prolific mouse on the planet and Linux can’t even use it, at all, natively.

On windows it just works.

Now let’s go deploy 300, or 3000 machines.

If your batteries are dying that fast, you’re doing something wrong.

You’re saying that data centers are replacing batteries constantly…just imagine the labor costs on that (and the down time), not even considering the material cost.

I work in enterprise, and have never heard such a thing. Even my friends in SMB routinely replace UPS’s at the 5-year mark, the same time they replace servers. They rarely replace batteries, at all, it’s so rare that it’s notable.

Well, what interfaces does it have?

How much space does it have inside?

How much power can it supply on the 12v bus?

Most likely the answer is “very little” or “not at all”.

I have a SFF box that can support, max, 3 SATA drives (2.5"), because of lack of ports and little space. But it has the power to run 5. So I put in a PCI SATA controller, and wedged them in. Runs fine.

I like the way you think.

This kind of approach solves so many problems, as the vendors have a vested interest