Host Setup: SUSE Virtualization Workshop¶

This workshop deploys on a bare metal Linux host with KVM. These steps cover host prep and initial deploy. Automation is entirely rodeo-cli; this repo ships the plan and the lab exercises. Same topology as suse-virt-rodeo.

No spare bare-metal host? See Deploy on AWS instead in the README. Same lab, an EC2 host does the work.

Requirements¶

baremetal/rodeo-plan.yaml ships one fixed sizing: 3× Harvester at 16 GiB/8 vCPU/320 GB, plus Rancher at 8 GiB/4 vCPU/60 GB. That's 56 GiB / 28 vCPU / 1020 GB guest total, and there's no separate tier system like AWS's budget/recommended/performance.

"Minimum" below is that config with host overhead. "Recommended" is the same config with comfortable margin for the host OS, buffer/cache, and not fighting for CPU during the cluster phase's nested-KVM install.

Resource	Minimum	Recommended
OS	SLES 16 / Leap 16 (Ubuntu 22.04+ and Fedora also work via `install-deps`)	same
RAM	64 GiB available	80-96 GiB available
vCPU	~32 free	~40+ free
Disk	~1050 GiB free in `/var/lib/libvirt/images` (320 GB × 3 Harvester + Rancher + ISO cache)	same, or more if you'll create many extra VMs during the exercises
KVM	`/dev/kvm` present. Nested virt only needed if the host itself is a VM	same
Network	Internet access during deploy	same

If your host has real headroom above the minimum, raise resources.harvester.memory_mib in rodeo-plan.yaml from 16384 toward 24576 (24 GiB/node) before your first rodeo up. That's the exact sizing the AWS variant uses once it has RAM to spare, live-verified to leave a comfortable margin instead of just barely fitting.

Confirm readiness:

rodeo doctor

You want a recommendation of harvester (or enough RAM to run it).

Install rodeo-cli¶

curl -fsSL https://raw.githubusercontent.com/avaleror/rodeo-cli/main/install.sh | bash

This clones the CLI, sets up its Python environment, and links rodeo as a system command. No venv to activate, no PATH tweaks, no sudo prefix on day-2 commands. rodeo up self-escalates when it needs root.

To pin a release:

curl -fsSL https://raw.githubusercontent.com/avaleror/rodeo-cli/main/install.sh | bash -s -- --ref v0.14.2

Deploy this workshop¶

git clone https://github.com/avaleror/suse-virt-workshop.git
cd suse-virt-workshop/baremetal
rodeo up

Because you are inside a directory that already contains rodeo-plan.yaml, rodeo up auto-detects the lab. It will:

Wrap itself in a tmux session so a dropped SSH does not kill the deploy
Re-run host checks and offer to install missing packages
Generate ~/.rodeo/secrets.yaml with random credentials
Drive the pipeline: kvm_host → vms → pxe_server → cluster → rancher → finalise
Print Harvester / Rancher URLs and where to find passwords

Typical time: 90-150 minutes. The cluster phase (iPXE install of Harvester inside nested KVM) is the long pole.

Survive disconnects¶

tmux attach -t rodeo-harvester    # session name matches the profile / lab
# Detach without stopping: Ctrl+b  d

Watch progress¶

From a second shell on the host:

rodeo watch                 # phases + serial consoles
rodeo logs harvester1       # one node's serial log
rodeo status                # VM state + VIP reachability

harvester_auto_import stays false¶

rodeo-plan.yaml sets harvester_auto_import: false on purpose. Importing Harvester into Rancher is Exercise 1. Do not import it before handing the lab to students.

Verify before students start¶

# All 3 Harvester nodes Ready
sudo rodeo ssh harvester1 -c 'export KUBECONFIG=/etc/rancher/rke2/rke2.yaml; sudo -E /var/lib/rancher/rke2/bin/kubectl get nodes'

# Rancher + Harvester APIs reachable
curl -sk https://192.168.122.9:30002/v3 | jq -r '.type'
curl -sk https://192.168.122.10/v1 | jq -r '.apiVersion'

# Only "local" in Rancher - Harvester not imported yet
curl -sk -u admin:$(grep rancher_admin_password ~/.rodeo/secrets.yaml | cut -d'"' -f2) \
  https://192.168.122.9:30002/v3/clusters | jq -r '.data[].name'

External DNAT (from a laptop on the same network):

curl -sk https://<host-ip>:8443/v1 | jq -r '.apiVersion'
curl -sk https://<host-ip>:30002/v3 | jq -r '.type'

See the pre-lab checklist for the full gate.

Hand students¶

Item	Value
Harvester UI	`https://<host-ip>:8443`
Rancher UI	`https://<host-ip>:30002`
Username	`admin`
Passwords	`harvester_admin_password` and `rancher_admin_password` in `~/.rodeo/secrets.yaml`
Lab guide	https://avaleror.github.io/suse-virt-workshop/

Both UIs use self-signed certificates. Students must accept the browser warning.

Tear down / redeploy¶

rodeo clean --yes                     # destroy this lab's VMs and disks
rodeo up                              # fresh deploy (reuses secrets)

rodeo clean --all --yes --secrets     # full host reset, wipe credentials
rodeo up                              # new secrets + fresh deploy

Resume a failed deploy¶

rodeo status
rodeo deploy --from cluster           # resume from the failed phase
# or
rodeo deploy --force                  # redo every phase

Optional: dedicated data disk¶

If the host has a second disk for lab images:

# in rodeo-plan.yaml
storage:
  device: /dev/nvme1n1          # verify with lsblk first
  mount_point: /var/lib/libvirt/images
  image_dir: /var/lib/libvirt/images
  fs_type: xfs

The kvm_host phase partitions, formats, and mounts it before creating VMs.