Used office PCs are some of the best value home servers you’ll come across. The Lenovo ThinkCentre, HP Elitedesk, and Dell Optiplex are fantastic machines with oodles of official documentation available straight from the OEM, and many come with built-in OOB management in the form of Intel AMT.

I have variants of all three. Love them.

Fair points. My entire homelab setup of five PCs pulls a total of 90-120W at any given time.

I’m gonna go check that 6th gen now that I’m home…

Intriguing, but not within the scope of this post. I’m not asking for KVM solutions.

I’m already very familiar with the AMT portion of vPro, all three of my Proxmox nodes have it enabled and locked down. Really handy to get in there when needed. The KVM route is rather expensive as I would need one that supports at least 5 systems.

vPro’s out-of-band management is the entire reason I use it, because my little lab is tucked in a utility room all the way in the basement, where I would have to cross the treacherous lands of scattered children’s toys.

I’ve already built the “new” NAS. Just trying to figure out the CPU situation before I take the plunge and swap the data drives over.

As for documentation, it really depends on the vendor, but the general process is the same overall. Here’s a PDF guide from MeshCentral that goes into more detail.

I use the CPU lists on Wikichip (Kaby Lake linked) to figure out what CPUs are compatible with vPro. Something to keep in mind is both the CPU and motherboard require vPro support for it to work properly.

Also a good point. Speaking of, that generation Optiplex SFF had a 300W PSU as an option in the XE3 variant (basically a 7050 meant for point-of-sale use) vs the stock 180W PSU. It’s plug-and-play, too. One of my Proxmox nodes runs a 7050 SFF with that PSU. It’s rock solid.

Yes, exactly. The shop was probably thinking of the cheap Molex ones.

I went to a local computer store and they were not very helpful. I asked if I could use a splitter for the power port and they said I would fry my board.

They aren’t wrong. Those SATA power splitters can be problematic due to subpar wiring and have been known to burn/melt.

Since it has come up a few times, in addition to the note in the git repository, I would like to clarify that XPipe is not fully FOSS software. The core that you can find on GitHub is Apache 2.0 licensed, but the distribution you download ships with closed-source extensions. There’s also a licensing system in place with limitations on what kind of systems you can connect to in the community edition as I am trying to make a living out of this. I understand that this is a deal-breaker for some, so I wanted to give a heads-up.

I appreciate the up-front attitude here, legitimately.

Sounds like it’s time to fire up another dedicated VM

Card games and board games with people. New Year’s isn’t really the time or place for this kind of thing.

Planka is fantastic kanban software. There is a 3rd party mobile app, but so far it hasn’t yet been updated to support Planka v2. Planka’s own mobile web UI is better than it used to be, but it’s not quite there yet.

The mobile app is 3rd party and has not yet been updated for Planka v2.

Source: I use Planka a LOT.

OpenMediaVault is based on Debian. I think it’s currently OMV 7, but I’m not at home at the moment so I can’t check. Very solid system though.

My NAS is a 2014 Mac Mini running OMV. It works great, very capable little Linux machine. Don’t bother with Mac OS.

For the file server conundrum, something to keep in mind is that Proxmox is not NAS software and isn’t really set up to do that kind of thing. Plus, the Proxmox devs have been very clear about not installing anything that isn’t absolutely necessary outside of Proxmox (on the same machine).

However, you can set up a file server inside an LXC and share that through an internal VLAN inside Proxmox. Just treat that LXC as a NAS.

For your *arr stack, fire up an exclusive VM just for them. Install Docker on the VM, too, of course.

LLMs

If you’re gonna use that, please make sure you comb through the output and understand it before implementing it.