I’m using Proxmox on a N100 16GB RAM & 512GB SSD. I’ve 5 VMs and LXCs (CTs) running: HA, ESPHome, CompreFace, Codeserver, Uptime Kuma. Total CPU usage is mostly at around 10%. The CPU load peaks are at 45%. Everything runs like a charm. The biggest benefit of Proxmox is that you can make snapshots and backups. This is so easy to do and especially snapshots are great with HA. Make a snapshot, install or change something and if it doesn’t work then rollback to previous state in seconds. USB passthrough also works fine. I use a Sonoff Zigbee Stick. I’ve no experience with Windows running on that PC. Maybe someone else can comment.
What I would recommend is that you purchase a 1TB SSD and upgrade your mini PC with it. If you use snapshots a lot then your 250 GB SSD is quickly becoming too small.
since you’re running so much, given not using much resources, just another VM with Win11 should be fine.
I wouldn’t need snapshots that much, just thought only HA on that box (native) is a bit of waste.
I can say that running a VM with Win11 is no problem, but GPU passthrough is currently a pain to setup and not supported for all scenarios. So do not expect your Win11 VM to do Video Encoding or 3D acceleration.
Also consider that N100 supports max 16GB RAM. Windows 11 needs at least 4GB - better 6 or 8 - this is half of your RAM only used by one VM!
HA would need another 4GB depending on your installed Addons etc. Besides HA you might want to setup some other lightweight VMs such as a reverse proxy or an OMV, which would need 1GB RAM each as well.
Hi,
Can you provide a link to your N100 pc please? I ordered a mini pc with n100 16gb and 512gb ssd but I was not able to install proxmox or even ubuntu.
depends, if you want to run local voice control in the future, going baremetal is recommended, else, if you do only automation and adding device on it, going VM is fine.
to be honest regardless whatever you workload is, if you want hassle free, no fuss, no complication, no headache, going baremetal is never a mistake.
HA is running on two cores, so the 4 other VMs are running on the remaining 2 cores? How does that work?
(My N100 hasn’t left China yet and I’ve never installed Proxmox)
I had some time to collect infos from my system. Here’s a screenshot. As you can see the CPU is mostly running at below 20% of its capacity. There are peaks every now and then. This happens when HA is updated or restarted but it is no issue for the other containers/VM’s.
Ok, great. So you give 2 cores to HA so that it will be ok doing intensive tasks like compiling an ESPNow firmware or whatnot but it will share these two cores with other VMs the rest of the time