I was bored. I’m not running it in a docker VM any more. I am now running it directly on FreeBSD using podman and FreeBSD’s Linux compatibility layer.

We don’t all live in overcrowded cities.

When do I get charged??

I’ve had Starlink for two years. My monthly bill is $55 a month, and no rental either…

Orrrr, maybe just let it sync with an ldap server.

What part of my comment makes you think I was promoting FreeBSD? I was replying to the part where you said people should use OPNsense under a VM because of ZFS.

People recommending against proxmox don’t understand how good zfs is

I was saying you can install OPNsense on bare metal and still use ZFS BECAUSE IT IS FreeBSD, and you simply select “ZFS” in the OpnSense installer to take advantage of ZFS without needing to add virtualization in front of system critical infrastructure.

So.

I did a thing.

I have audiomuse-ai running its main, complete docker compose script, with all containers, on my 8GB Raspi5, and worker-only containers running on:

• An 8GB bhyve VM on my FreeBSD box
• An E2-6110 AMD pre-ryzen APU with 16GB of ddr3
• A Ryzen 5800x w 32GB RAM

They’ve been running about a week, and I’m a little over a third of the way through

Once the initial analysis is complete, I’ll stop all worker containers and leave it all just running fully on the pi5.

I also created a worker-only addon for the 6600T machine, but as it is already running HAOS and Jellyfin, I was getting a lot of OOM-related failures when it was running.

But I also have 32G of used, eBay bought, ddr4 SODIMMs.coming for it.

Bonus: Most of my homelab is in this. The only things missing are my Sophos running OPNsense, and the raspi5. Oh, and my actual desktop machine.

Mine, a Brocade ICX6450, is solid as a rock using a hacked license to enable 10Gbe on all 4 SFP+ ports.

They’re loud, as the firmware is a little buggy, and doesn’t ever actually set the fan speed to low after booting, regardless of the temps.

But there’s a few tricks to that. There’s a command you can run to force the speed to low, and I just have that sent via SSH when my raspi5 boots. Because it needs to be sent when the switch powers on. So if I lose power for an extended time, it’ll still be reset to low after it all comes back.

My OPNsense setup is on bare metal. It’s a Sophos SG135 rev 2 with 6GB of RAM and a 64GB NVme SSD.

It can be upgraded to 16GB, but isn’t nice for my set up.

I don’t use Proxmox, but I do make extensive use of ZFS across most of my entire homelab.

My NAS/Media server has 48T of spinning SAS3 drives, runs FreeBSD 15.1, and has a BhyVE VM running Alpine Linux and docker for the 1 or 2 services I use that simply won’t run easily on FreeBSD.

I run most of the rest of my services in jails on that host, jails are what linux’s entire container subsystem is based on, having been around for 26 years now. Yes, FreeBSD’s jail system was introduced in 2000.

I have a raspi 5 running rasbian, with Adguard Home, and audiomuse-ai on it.

And a Lenovo M700 Tiny running Home Assistant.

Tying it all together is a managed brocade/ruckus switch in layer 3 routing mode, handling all routing, VLANs, subnets, etc…

I had a Linux box with two 10Mbps NICs in it in the mid 90s to early 2000s doing NAT so I could share the cable modem connection to my wife’s computer back when you were only allowed to have a single machine connected to the Internet at home.

I say all that to lay out my experience level.

With all that said, you can virtualize your primary router if you like. Personally, I’d rather that system critical piece of equipment be fully isolated from any possible virtualization shenanigans.

Not to mention what happens when you fiddle with your Proxmox setup too much and oops, you have no Internet now.

What happens when your main network goes down, and the only way you can access that Proxmox machine is over that network?

Or maybe they do understand how good ZFS is, and since OPNsense is FreeBSD based, they use ZFS IN OPNsense.

OPNsense can make snapshots and restore them native from right inside the UI.

Which someone who used it should know.

I’ve been considering audiomuse, but I have old equipment available.

My options are my media server, which is an old Xeon E3-1275v3 with 32G of RAM, which also hosts Navidrome, my arr stack and the associated downloaders, or my Home Assistant and Jellyfin box, which is a Lenovo M700 Tiny which is an i5 6600T but has only 8G of ram.

Or, an 8G Pi5 with an SSD (using the pi SSD hat)

I’m not sure either of those 3 options would handle audiomuse AI all that well…