As someone who moved to Proxmox for my 3-node homelab, good luck.
I find the automation for deploying VMs to be woefully incapable compared to Terraform/PowerCLI on the VMware side. Not to mention things like load balancing/DRS are flat out missing.
I managed to get it stable enough for homelab-y things like *arr, plex, DNS, etc - but at this point I would quit rather than use it in a production environment. Or maybe I would just look at bare metal kubernetes instead.
The biggest issue is being in aware of migrations for load balancing. If VM 1 is deployed to Node 1 with Terraform, then is moved to Node 2 at some point for load balancing, Terraform tries to recreate it on Node 1.
Also, I have a slight moral objection to one of the top providers being developed by a for-profit prison company.
IaaS or gtfo? I would love to see more development in this area, but I think you might be covering a bit too much ground with “in a production environment”. Tons of smaller (and not so small) companies are still running piles of bare metal chaos and could benefit greatly from even the simplest Proxmox setup.
As someone who moved to Proxmox for my 3-node homelab, good luck.
I find the automation for deploying VMs to be woefully incapable compared to Terraform/PowerCLI on the VMware side. Not to mention things like load balancing/DRS are flat out missing.
I managed to get it stable enough for homelab-y things like *arr, plex, DNS, etc - but at this point I would quit rather than use it in a production environment. Or maybe I would just look at bare metal kubernetes instead.
Have you seen xcp-ng and xen orchestra?
What OS would you use for the bare metal install?
Probably Debian or Ubuntu LTS?
Your use case sounds like kubernetes would be a way better fit as dynamicly scaling and load balancing is kinda the whole point of kubernetes.
Proxmox clustering is essentially just for adding redundancy and nothing more.
Huh, I use terraform for my Proxmox clusters without any major issues. What kind of trouble does it give you?
The biggest issue is being in aware of migrations for load balancing. If VM 1 is deployed to Node 1 with Terraform, then is moved to Node 2 at some point for load balancing, Terraform tries to recreate it on Node 1.
Also, I have a slight moral objection to one of the top providers being developed by a for-profit prison company.
IaaS or gtfo? I would love to see more development in this area, but I think you might be covering a bit too much ground with “in a production environment”. Tons of smaller (and not so small) companies are still running piles of bare metal chaos and could benefit greatly from even the simplest Proxmox setup.