govirt

RedHat has abandoned RHV and thusly oVirt is likely dead. This project will not be maintained going forward :(

About

govirt allows you to perform basic tasks in oVirt on the command line.

I highly recommend starting with --help before following the example below. There are very simple things you'll want to know how to set as you go, such as:

• the oVirt engine API URL
• engine authentication
• template name
• template version
• storage domain name

Example of building a 5-node kubernetes cluster using k3s

1. clean up previous tests (if any):

   # clear known_hosts
   for i in k3s{0..4} k3s{0..4}.domain 192.168.0.20{0..4}; do \
       ssh-keygen -R ${i}; \
   done;
   

   # force-stop any existing VMs with the same names
   for i in {0..4}; do \
       govirt vm stop \
           --force \
           --name "k3s${i}"; \
       done;
   
   # delete any existing VMs with the same names
   for i in {0..4}; do \
       govirt vm rm \
           --yes \
           --name "k3s${i}"; \
       done;
   

2. create new VMs:

   1. clone each VM from a sealed template:

      for i in {0..4}; do \
          govirt vm create \
              --name "k3s${i}" \
              --memory 4 \
              --cpu 2; \
          done;
      

   2. prepare cloud-init scripts for each new VM to run at bootup:

      mkdir -p "${HOME}/cloud-init" && \
      for i in {0..4}; do \
      govirt cloud-init create \
          --fqdn "k3s${i}.lan" \
          --ssh-key "ssh-rsa some_ssh_public_key with_comment_if_you_like" > "${HOME}/cloud-init/k3s${i}.yml"; \
      done;
      
      

   3. start VMs with cloud-init script:

      for i in {0..4}; do \
          govirt vm start \
              --name "k3s${i}" \
              --init \
              --script "${HOME}/cloud-init/k3s${i}.yml"; \
          done;
      

3. setup ansible controller:

   # wait for VMs to come up with correct hostname, ip address, etc and accept host ssh key
   for i in k3s{0..4}.domain; do \
       ssh -o StrictHostKeyChecking=accept-new "${i}" -t hostname -s; \
   done;
   

4. config management:

   # deploy config management and the k3s cluster
   ansible-playbook -i ~/ansible/inventory/kubernetes.yml -e reboot=yes ~/ansible/playbooks/k3s.yml
   
   # copy the kubeconfig from first node to controller node to make life easy
   rsync -avhP k3s0:"~/.kube/config" ~/.kube/config && sed -i 's|https://127.0.0.1|https://k3s0.domain|g' ~/.kube/config
   

5. verify cluster:

   kubectl get nodes -o wide
   
   kubectl get pods -A