Skip to content

SUSE Virtualization Workshop: Lab Guide

Version: 2.0 | Duration: ~3 hours (after deploy)
Author: Andres Valero, Principal Technology Advocate, SUSE

Self-hosted twin of suse-virt-rodeo. Same eight exercises (plus one self-hosted-only bonus), deployed with rodeo-cli.

Versions: Harvester 1.8.1 · Rancher 2.14.1 · K3s v1.35.3+k3s1 · rodeo-cli v0.14.x

0. Deploy the lab (instructor / self-serve)

curl -fsSL https://raw.githubusercontent.com/avaleror/rodeo-cli/main/install.sh | bash
git clone https://github.com/avaleror/suse-virt-workshop.git
cd suse-virt-workshop/baremetal   # or aws/, see README for the AWS path
rodeo doctor
rodeo up
  • Time: 90-150 min. Survive disconnects: tmux attach -t rodeo-harvester
  • UIs: Harvester https://<host-ip>:8443 · Rancher https://<host-ip>:30002
  • User admin · passwords on your claim-portal lab page, or in ~/.rodeo/secrets.yaml if you deployed the lab
  • Tear down: rodeo clean --yes

Full detail: Host setup.

Exercise 1: The Arrival (30 min)

  1. curl -sk https://192.168.122.10/v1 and :30002/v3 - both healthy.
  2. Rancher → Virtualization Management → Import Existing → name harvester → copy URL.
  3. Harvester → Settings → cluster-registration-url → paste → Save → wait until Active.
  4. Tour Dashboard / Hosts / Support (download kubeconfig).
  5. Save the kubeconfig as ~/.kube/config on the KVM host, point its server at https://192.168.122.10 → kubectl get nodes - three Ready.

Exercise 2: The Subterranean Divide (30 min)

  1. Inspect Hosts; kubectl get replicas.longhorn.io -n longhorn-system shows each volume's replicas per node.
  2. Create namespaces prod and dev.
  3. Review StorageClass harvester-longhorn (3 replicas).
  4. Create StorageClass harvester-longhorn-1rep (1 replica).
  5. Create VM network prod/service (UntaggedNetwork on mgmt).
  6. Install the lab VM key from /opt/suse-virt-workshop/lab-vm-key as ~/.ssh/id_ed25519, register its .pub as SSH key prod/default.
  7. Upload cloud image into official-images (wait Active).
  8. Optional: cloud-init user-data template prod/prod with qemu-guest-agent.

Exercise 3: The Flash Crash (30 min)

Create VM prod/algo-trader-01: 2 CPU / 2 GiB, SSH prod/default, label stage=prod, root disk from cloud image (5 GiB) + volume market-data-vol (1 GiB), network prod/service, network-data static IP 192.168.122.50/24. Console + ssh sles@192.168.122.50.

Exercise 4: The Rising Tide (25 min)

  1. webserver-prod and daily-batch-processor are already running on prod/service (deploy-time automation). Confirm both and note webserver-prod's node.
  2. Pause the batch VM; ping the gateway IP; leave it running.
  3. Migrate webserver-prod to another node; confirm ping/uptime survive.
  4. Unpause batch; put the old node in Maintenance Mode; watch evacuation.

Exercise 5: The Invisible Intruder (30 min)

  1. Cluster network closed-loop with uplink NIC (often ens5).
  2. VM network prod/secure-loop-prod (Untagged on closed-loop).
  3. Overlay dev/secure-loop-dev + subnet secure-vpc-dev (192.168.32.0/24).
  4. Edit a VM → attach isolated network → stop/start to apply.

Exercise 6: The Unthinkable Error (30 min)

  1. On transaction-ledger (create if needed): write ledger.txt, snapshot pre-disaster-backup, delete the file.
  2. Restore snapshot to new VM ledger-staging-verify; verify file.
  3. Power off prod → Restore snapshot → power on → verify.
  4. Attach the backup-target NFS setting to 192.168.122.1:/srv/backups/. Already exported by deploy-time automation, no host setup needed.
  5. Create a recurring backup/snapshot schedule.

Exercise 7: The Stampede (25 min)

  1. Template harvester-public/prod-basic (1 CPU / 1 GiB, prod/default, cloud image, prod/service, label stage=prod).
  2. Create calc-engine-01…03 from template; scale to five; delete extras.

Exercise 8: A New Horizon (10 min)

Review the skill table; bookmark SUSE Virtualization docs; tear down with rodeo clean --yes when finished.

Bonus: The Final Showdown (25 min, self-hosted only, no rodeo counterpart)

  1. Open Advanced → Migration → Sources (empty unless you point it at a real legacy hypervisor. Learn the form.)
  2. Create stand-in legacy-ledger-vm on prod/service.
  3. systemctl enable --now qemu-guest-agent inside the guest.
  4. Snapshot post-migration-baseline; live-migrate once.

Cheat sheet

Item Value
Harvester https://<host-ip>:8443
Rancher https://<host-ip>:30002
VIP / nodes .10 / .11–.13 · rancher .9
Passwords claim-portal lab page, or ~/.rodeo/secrets.yaml
Guest example ssh sles@192.168.122.50
Self-checks /opt/suse-virt-workshop/checks/check-exercise-N.sh
Ops (if you deployed the lab) rodeo status · sudo rodeo ssh harvester1 · rodeo watch