Possible to use control plane nodes for longhorn? Longhorn memory usage?

[TimHeckel]

Hi again :) I'm spinning up a new cluster in prep for migration; things are going well, but I noticed a couple things:

(1) Is it possible to specify which nodes longhorn chooses for disk allocation? It seems like it only chose the 3 agent nodes I set in kube.tf:

FWIW, my kube.tf nodepools config looks like:

control_plane_nodepools = [
    {
      name        = "sb-control-plane",
      server_type = "ccx22",
      location    = "ash",
      labels      = [],
      taints      = [],
      count       = 3
    }
  ]

  agent_nodepools = [
    {
      name        = "agent-small",
      server_type = "cpx11",
      location    = "ash",
      labels      = [],
      taints      = [],
      count       = 3
    }
  ]

(2) Per lens, with a vanilla install with longhorn enabled, my cluster memory usage was higher than I expected. Not sure if this is known when leveraging longhorn?

Thanks so much :)

[TimHeckel]

Oh. I mistakenly swapped the hetzner types between my control plane and my agents -- doh. cpx11 should have been set for the agents, and ccx22 for the agents! Then the longhorn UI would be correct (displaying only agents, not control plane nodes) and the memory will look MUCH better :) Sorry for the confusion.