Launching And Managing Instances on OpenShift (OKD) Virtualization with Ansible

Hello, this is me, international best-selling-au-- err, I mean, your friendly technologist. Today, I will be spinning instances on OpenShift Virtualization using Ansible. While there is already enough automation with OpenShift, Ansible brings with it, certain advantages that makes it an ideal tool for spinning OpenShift Virtualization instances.

Prerequisites

I will be using virtual-machine-compute to launch my instances. I already have my SSH public key stored in a secret in my environment, which is created as follows:

oc create secret generic example-secret-public-key --from-file=key1=my-key.pub -n virtual-machine-compute

I also need to set So I revised it:up external networking via nmstate, which is already done and you can read about here.

Finally, I need to make sure that I am authenicated. For this exercise, I just authenicated via my web browser with:

oc login --web

With that, it is time to install Ansible and Ansible/Kubernetes support. Note that I will be using uv for remainder of the walkthrough.

Installation

The first thing is to initialize my python environment:

➜  kubevirt git:(main) ✗ uv venv                  
Using CPython 3.11.14
Creating virtual environment at: .venv
Activate with: source .venv/bin/activate

Next I install ansible:

➜  ~ uv pip install ansible               
Resolved 22 packages in 28ms
Installed 9 packages in 246ms
 + ansible==12.3.0
 + ansible-core==2.19.6
 + cffi==2.0.0
 + cryptography==46.0.5
 + jinja2==3.1.6
 + markupsafe==3.0.3
 + packaging==26.0
 + pycparser==3.0
 + resolvelib==1.2.1

Then install the kubernetes dependencies:

➜  ~ uv pip install kubernetes
Resolved 13 packages in 68ms
Prepared 1 package in 51ms
Installed 13 packages in 22ms
 + certifi==2026.1.4
 + charset-normalizer==3.4.4
 + durationpy==0.10
 + idna==3.11
 + kubernetes==35.0.0
 + oauthlib==3.3.1
 + python-dateutil==2.9.0.post0
 + pyyaml==6.0.3
 + requests==2.32.5
 + requests-oauthlib==2.0.0
 + six==1.17.0
 + urllib3==2.6.3
 + websocket-client==1.9.0

Finally, I install via ansible-galaxy the kubevirt.core collection, which will also install the following:

  • kubernetes.core
  • community.okd

Behind the scenes:

➜  ~ uv run ansible-galaxy collection install kubevirt.core
Starting galaxy collection install process
Process install dependency map
Starting collection install process
Downloading https://galaxy.ansible.com/api/v3/plugin/ansible/content/published/collections/artifacts/kubevirt-core-2.2.3.tar.gz to /Users/rilindo/.ansible/tmp/ansible-local-68764maasu3v3/tmpz8xo97ww/kubevirt-core-2.2.3-ieyravzl
Installing 'kubevirt.core:2.2.3' to '/Users/rilindo/.ansible/collections/ansible_collections/kubevirt/core'
Downloading https://galaxy.ansible.com/api/v3/plugin/ansible/content/published/collections/artifacts/kubernetes-core-6.3.0.tar.gz to /Users/rilindo/.ansible/tmp/ansible-local-68764maasu3v3/tmpz8xo97ww/kubernetes-core-6.3.0-66plso9b
kubevirt.core:2.2.3 was installed successfully
Installing 'kubernetes.core:6.3.0' to '/Users/rilindo/.ansible/collections/ansible_collections/kubernetes/core'
kubernetes.core:6.3.0 was installed successfully

Next, I create directory ~/kubevirt to store my ansible code. I then change over to that directory and generate the ansible.cfg file

uv run ansible-config init --disabled > ansible.cfg

In the ansible.cfg, I updated inventory under [defaults]

 [defaults]
 inventory=inventory/

Then I created the directory and in that directory, I created the file kubevirt.yaml

plugin: kubevirt.core.kubevirt
namespaces:
  - virtual-machine-compute
network_name: nic-white-chimpanzee-56

This inventory file initializes KubeVirt and pulls the list of instances in the virtual-machine-compute namespace. The network name is the external interface attached to each instance, which Ansible will use to connect.

With that setup, I am able to bring up the list of hosts in that namespace:

➜  kubevirt git:(main) ✗ uv run ansible --list-hosts all
  hosts (5):
    virtual-machine-compute-freebsd-14-8giusc6ap5mnlven
    virtual-machine-compute-rocky-vm-ex188
    virtual-machine-compute-rocky-vm-lb
    virtual-machine-compute-rocky-vm-nmstate
    virtual-machine-compute-rocky-vm-sr

Now I can proceed to spin up a single instance.

Spinning Up A Single Instance

Previously, when I spin up an instance, I used the following manifest to create it:

---
apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: rocky-dv-nmstate
  namespace: virtual-machine-compute
spec:
  source:
    pvc:
      name: rockylinux-10-iscsi
      namespace: custom-vm-images
  pvc:
    accessModes:
      - ReadWriteMany
    resources:
      requests:
        storage: 20Gi
    storageClassName: synology-iscsi-storage
    volumeMode: Block

---
apiVersion: kubevirt.io/v1
kind: VirtualMachine
metadata:
  name: rocky-vm-nmstate
  namespace: virtual-machine-compute
spec:
  runStrategy: RerunOnFailure
  template:
    metadata:
      labels:
        kubevirt.io/domain: rocky-vm-nmstate
        app: webserver-nmstate
    spec:
      evictionStrategy: LiveMigrate
      accessCredentials:
        - sshPublicKey:
            propagationMethod:
              noCloud: {}
            source:
              secret:
                secretName: example-secret-public-key
      domain:
        cpu:
          cores: 1
          sockets: 1
          threads: 1
        memory:
          guest: 4Gi
        devices:
          disks:
            - disk:
                bus: virtio
              name: rootdisk
            - disk:
                bus: virtio
              name: cloudinitdisk
          interfaces:
            - masquerade: {}
              model: virtio
              name: default
            - bridge: {}
              model: virtio
              name: nic-white-chimpanzee-56
              state: up
      networks:
        - name: default
          pod: {}
        - multus:
            networkName: nad-efficient-grouse
          name: nic-white-chimpanzee-56
      volumes:
        - name: rootdisk
          dataVolume:
            name: rocky-dv-nmstate
        - cloudInitNoCloud:
            networkData: |
              version: 2
              ethernets:
                enp1s0:
                  dhcp4: true
                  # Higher metric means lower priority for the pod network
                  dhcp4-overrides:
                    route-metric: 200
                enp2s0:
                  dhcp4: true
                  # Lower metric means this becomes the preferred default route
                  dhcp4-overrides:
                    route-metric: 50
                  # Explicit static route for your workstation network
                  routes:
                    - to: 192.168.1.0/24
                      via: 192.168.3.1
            userData: |-
              #cloud-config
              user: cloud-user
              chpasswd: { expire: False }
              packages:
                - httpd
              runcmd:
                - "sudo systemctl enable httpd --now"
                - "echo '<html><body><h1>Rocky Linux Web Server on KubeVirt here!</h1></body></html>' | sudo tee /var/www/html/index.html"
          name: cloudinitdisk

I took that manifest and adapted to the following playbook.

- hosts: localhost
  connection: local
  tasks:
  - name: Create VM
    kubevirt.core.kubevirt_vm:
      state: present
      wait: true
      name: rockylinux-built-with-ansible
      namespace: virtual-machine-compute
      labels:
        app: rockylinux-built-with-ansible
      data_volume_templates:
        - metadata:
            name: rockylinux-built-with-ansible
          spec:
            source:
              pvc:
                name: rockylinux-10-iscsi
                namespace: custom-vm-images
            pvc:
              accessModes:
                - ReadWriteMany
              resources:
                requests:
                  storage: 20Gi
              storageClassName: synology-iscsi-storage
              volumeMode: Block
      spec:
        evictionStrategy: LiveMigrate
        accessCredentials:
          - sshPublicKey:
              propagationMethod:
                noCloud: {}
              source:
                secret:
                  secretName: example-secret-public-key
        domain:
          cpu:
            cores: 1
            sockets: 1
            threads: 1
          memory:
            guest: 4Gi
          devices:
            disks:
              - disk:
                  bus: virtio
                name: rootdisk
              - disk:
                  bus: virtio
                name: cloudinitdisk
            interfaces:
              - masquerade: {}
                model: virtio
                name: default
              - bridge: {}
                model: virtio
                name: nic-white-chimpanzee-56
                state: up
        networks:
          - name: default
            pod: {}
          - multus:
              networkName: nad-efficient-grouse
            name: nic-white-chimpanzee-56
        volumes:
          - name: rootdisk
            dataVolume:
              name: rockylinux-built-with-ansible
          - cloudInitNoCloud:
              networkData: |
                version: 2
                ethernets:
                  enp1s0:
                    dhcp4: true
                    # Higher metric means lower priority for the pod network
                    dhcp4-overrides:
                      route-metric: 200
                  enp2s0:
                    dhcp4: true
                    # Lower metric means this becomes the preferred default route
                    dhcp4-overrides:
                      route-metric: 50
                    # Explicit static route for your workstation network
                    routes:
                      - to: 192.168.1.0/24
                        via: 192.168.3.1
              userData: |-
                #cloud-config
                user: cloud-user
                chpasswd: { expire: False }
                packages:
                  - httpd
                runcmd:
                  - "sudo systemctl enable httpd --now"
                  - "echo '<html><body><h1>Rocky Linux Web Server on KubeVirt here!</h1></body></html>' | sudo tee /var/www/html/index.html"
            name: cloudinitdisk

As you can see, it isn't much different from a Kubernetes template. The key difference are mostly the location of labels, names, alongside the inclusion of Ansible attributes (such as state: present) as well as using DataVolume template for persistent storage.

I run the playbook:

➜  kubevirt git:(main) ✗ uv run ansible-playbook single-vm.yaml

PLAY [localhost] ***************************************************************************************************************************************************************************************

TASK [Gathering Facts] *********************************************************************************************************************************************************************************
ok: [localhost]

TASK [Create VM] ***************************************************************************************************************************************************************************************
changed: [localhost]

PLAY RECAP *********************************************************************************************************************************************************************************************
localhost                  : ok=2    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

And I can see that it launched

➜  kubevirt git:(main) ✗ uv run ansible --list-hosts all
  hosts (6):
    virtual-machine-compute-freebsd-14-8giusc6ap5mnlven
    virtual-machine-compute-rocky-vm-ex188
    virtual-machine-compute-rocky-vm-lb
    virtual-machine-compute-rocky-vm-nmstate
    virtual-machine-compute-rocky-vm-sr
    virtual-machine-compute-rockylinux-built-with-ansible

However, I only care about the instances I launch with Ansible, so I made a small adjustment to the inventory file so that only instances with the label rockylinux-built-with-ansible are listed:

plugin: kubevirt.core.kubevirt
namespaces:
  - virtual-machine-compute
network_name: nic-white-chimpanzee-56
label_selector: app=rockylinux-built-with-ansible

Which should return this:

➜  kubevirt git:(main) ✗ uv run ansible --list-hosts all
  hosts (1):
    virtual-machine-compute-rockylinux-built-with-ansible

With that, I can verify that the instance is running:

➜  kubevirt git:(main) ✗ uv run ansible -m ping all -u cloud-user
[WARNING]: Host 'virtual-machine-compute-rockylinux-built-with-ansible' is using the discovered Python interpreter at '/usr/bin/python3.12', but future installation of another Python interpreter could cause a different interpreter to be discovered. See https://docs.ansible.com/ansible-core/2.19/reference_appendices/interpreter_discovery.html for more information.
virtual-machine-compute-rockylinux-built-with-ansible | SUCCESS => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/bin/python3.12"
    },
    "changed": false,
    "ping": "pong"
}

And I can log into the instance:

➜  kubevirt git:(main) ✗ ssh cloud-user@rockylinux-built-with-ansible
The authenticity of host 'rockylinux-built-with-ansible (192.168.3.118)' can't be established.
ED25519 key fingerprint is SHA256:cqFLHf42TbqZpMwpixNapfj0hMyNXDDz4oQUgQoZBtU.
This key is not known by any other names.
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added 'rockylinux-built-with-ansible' (ED25519) to the list of known hosts.
[cloud-user@rockylinux-built-with-ansible ~]$ uptime
 18:11:23 up 1 min,  2 users,  load average: 0.88, 0.29, 0.10
[cloud-user@rockylinux-built-with-ansible ~]$

Looking good. Now I will spin up multiple instances.

Spinning Up Multiple Instances

I created a new playbook that spins up three instances:

---
- name: Create Virtual Machine Instances
  hosts: localhost
  connection: local
  vars:
    vm_list:
      - rocky01-instance
      - rocky02-instance
      - rocky03-instance
  tasks:
  - name: Create/Destroy VM instances
    loop: "{{ vm_list }}"
    kubevirt.core.kubevirt_vm:
      state: "{{ vm_state }}"
      wait: true
      name: "{{ item }}"
      namespace: virtual-machine-compute
      labels:
        app: rockylinux
      data_volume_templates:
        - metadata:
            name: "{{ item }}"
          spec:
            source:
              pvc:
                name: rockylinux-10-iscsi
                namespace: custom-vm-images
            pvc:
              accessModes:
                - ReadWriteMany
              resources:
                requests:
                  storage: 20Gi
              storageClassName: synology-iscsi-storage
              volumeMode: Block
      spec:
        evictionStrategy: LiveMigrate
        accessCredentials:
          - sshPublicKey:
              propagationMethod:
                noCloud: {}
              source:
                secret:
                  secretName: example-secret-public-key
        domain:
          cpu:
            cores: 1
            sockets: 1
            threads: 1
          memory:
            guest: 4Gi
          devices:
            disks:
              - disk:
                  bus: virtio
                name: rootdisk
              - disk:
                  bus: virtio
                name: cloudinitdisk
            interfaces:
              - masquerade: {}
                model: virtio
                name: default
              - bridge: {}
                model: virtio
                name: nic-white-chimpanzee-56
                state: up
        networks:
          - name: default
            pod: {}
          - multus:
              networkName: nad-efficient-grouse
            name: nic-white-chimpanzee-56
        volumes:
          - name: rootdisk
            dataVolume:
              name: "{{ item }}"
          - cloudInitNoCloud:
              networkData: |
                version: 2
                ethernets:
                  enp1s0:
                    dhcp4: true
                    # Higher metric means lower priority for the pod network
                    dhcp4-overrides:
                      route-metric: 200
                  enp2s0:
                    dhcp4: true
                    # Lower metric means this becomes the preferred default route
                    dhcp4-overrides:
                      route-metric: 50
                    # Explicit static route for your workstation network
                    routes:
                      - to: 192.168.1.0/24
                        via: 192.168.3.1
              userData: |-
                #cloud-config
                user: cloud-user
                chpasswd: { expire: False }
            name: cloudinitdisk

This is where the advantages of Ansible become apparent, as I can parameterize specific values in the manifest. In this case, I generate a list of instances in a list value vm_list, then I loop through the list to create distinct instances and associate storage using {{ item }}. Those instances, though, will only be created as long as the vm_state is defined. This means less duplicate code as well as avoiding somewhat complex tooling to accomplish the same thing.

At this point, since I updated the label to rockylinux, I just need to update the inventory:

plugin: kubevirt.core.kubevirt
namespaces:
  - virtual-machine-compute
network_name: nic-white-chimpanzee-56
label_selector: app=rockylinux

And then I run the playbook:

➜  kubevirt git:(main) ✗ uv run ansible-playbook multiple-vm.yaml -e vm_state=present
[WARNING]: provided hosts list is empty, only localhost is available. Note that the implicit localhost does not match 'all'

PLAY [Create Virtual Machine Instances] ****************************************************************************************************************************************************************

TASK [Gathering Facts] *********************************************************************************************************************************************************************************
ok: [localhost]

TASK [Create/Destroy VM instances] *********************************************************************************************************************************************************************
changed: [localhost] => (item=rocky01-instance)
changed: [localhost] => (item=rocky02-instance)
changed: [localhost] => (item=rocky03-instance)

PLAY RECAP *********************************************************************************************************************************************************************************************
localhost                  : ok=2    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

Configuring Instances

Of course, I do need to configure the instances, which is Ansible's core strength. So I whip up the following playbook multiple-vm-configure.yaml to get the hostname and uptime, install nginx and start it up.

---
- name: Configure Virtual Machine Instances
  hosts: namespace_virtual_machine_compute
  become: yes
  remote_user: cloud-user
  tasks:
  - name: Display uptime
    ansible.builtin.debug:
      msg: "Uptime for {{ inventory_hostname }}: {{ now().replace(microsecond=0) - now().fromtimestamp(now(fmt='%s') | int - ansible_uptime_seconds) }}"
  - name: Install nginx
    ansible.builtin.dnf:
      name: nginx
      state: latest
  - name: Start up nginx
    ansible.builtin.systemd_service:
      name: nginx
      state: started
      enabled: true

This will get the uptime, install Nginx and then start it up.

I then ran, and confirmed it was successful:

➜  kubevirt git:(main) ✗ uv run ansible-playbook  multiple-vm-configure.yaml

PLAY [Configure Virtual Machine Instances] *************************************************************************************************************************************************************

TASK [Gathering Facts] *********************************************************************************************************************************************************************************
[WARNING]: Host 'virtual-machine-compute-rocky02-instance' is using the discovered Python interpreter at '/usr/bin/python3.12', but future installation of another Python interpreter could cause a different interpreter to be discovered. See https://docs.ansible.com/ansible-core/2.19/reference_appendices/interpreter_discovery.html for more information.
ok: [virtual-machine-compute-rocky02-instance]
[WARNING]: Host 'virtual-machine-compute-rocky01-instance' is using the discovered Python interpreter at '/usr/bin/python3.12', but future installation of another Python interpreter could cause a different interpreter to be discovered. See https://docs.ansible.com/ansible-core/2.19/reference_appendices/interpreter_discovery.html for more information.
ok: [virtual-machine-compute-rocky01-instance]
[WARNING]: Host 'virtual-machine-compute-rocky03-instance' is using the discovered Python interpreter at '/usr/bin/python3.12', but future installation of another Python interpreter could cause a different interpreter to be discovered. See https://docs.ansible.com/ansible-core/2.19/reference_appendices/interpreter_discovery.html for more information.
ok: [virtual-machine-compute-rocky03-instance]

TASK [Display uptime] **********************************************************************************************************************************************************************************
ok: [virtual-machine-compute-rocky01-instance] => {
    "msg": "Uptime for virtual-machine-compute-rocky01-instance: 0:02:32"
}
ok: [virtual-machine-compute-rocky02-instance] => {
    "msg": "Uptime for virtual-machine-compute-rocky02-instance: 0:01:30"
}
ok: [virtual-machine-compute-rocky03-instance] => {
    "msg": "Uptime for virtual-machine-compute-rocky03-instance: 0:00:27"
}

TASK [Install nginx] ***********************************************************************************************************************************************************************************
changed: [virtual-machine-compute-rocky01-instance]
changed: [virtual-machine-compute-rocky03-instance]
changed: [virtual-machine-compute-rocky02-instance]

TASK [Start up nginx] **********************************************************************************************************************************************************************************
changed: [virtual-machine-compute-rocky03-instance]
changed: [virtual-machine-compute-rocky02-instance]
changed: [virtual-machine-compute-rocky01-instance]

PLAY RECAP *********************************************************************************************************************************************************************************************
localhost                  : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
virtual-machine-compute-rocky01-instance : ok=4    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
virtual-machine-compute-rocky02-instance : ok=4    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
virtual-machine-compute-rocky03-instance : ok=4    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   

➜  kubevirt git:(main) ✗

Then I made some updates - I don't want a generic nginx page, but a customized page with the hostname and the point in time snapshot of the uptime. So I revised it to:

---
- name: Configure Virtual Machine Instances
  hosts: namespace_virtual_machine_compute
  become: yes
  remote_user: cloud-user
  vars: 
    nginx_default_page: /usr/share/nginx/html/index.html
  tasks:
  - name: Display uptime
    ansible.builtin.debug:
      msg: "Uptime for {{ inventory_hostname }}: {{ now().replace(microsecond=0) - now().fromtimestamp(now(fmt='%s') | int - ansible_uptime_seconds) }}"
  - name: Install nginx
    ansible.builtin.dnf:
      name: nginx
      state: latest
  - name: Start up nginx
    ansible.builtin.systemd_service:
      name: nginx
      state: started
      enabled: true
  - name: Generate default page
    ansible.builtin.copy:
      content: "Uptime for {{ inventory_hostname }}: {{ now().replace(microsecond=0) - now().fromtimestamp(now(fmt='%s') | int - ansible_uptime_seconds) }}"
      dest: "{{ nginx_default_page }}"
      mode: 0644
      owner: root
      group: root

And re-ran it:

➜  kubevirt git:(main) ✗ uv run ansible-playbook multiple-vm-configure.yaml

PLAY [Configure Virtual Machine Instances] *************************************************************************************************************************************************************

TASK [Gathering Facts] *********************************************************************************************************************************************************************************
[WARNING]: Host 'virtual-machine-compute-rocky02-instance' is using the discovered Python interpreter at '/usr/bin/python3.12', but future installation of another Python interpreter could cause a different interpreter to be discovered. See https://docs.ansible.com/ansible-core/2.19/reference_appendices/interpreter_discovery.html for more information.
ok: [virtual-machine-compute-rocky02-instance]
[WARNING]: Host 'virtual-machine-compute-rocky03-instance' is using the discovered Python interpreter at '/usr/bin/python3.12', but future installation of another Python interpreter could cause a different interpreter to be discovered. See https://docs.ansible.com/ansible-core/2.19/reference_appendices/interpreter_discovery.html for more information.
ok: [virtual-machine-compute-rocky03-instance]
[WARNING]: Host 'virtual-machine-compute-rocky01-instance' is using the discovered Python interpreter at '/usr/bin/python3.12', but future installation of another Python interpreter could cause a different interpreter to be discovered. See https://docs.ansible.com/ansible-core/2.19/reference_appendices/interpreter_discovery.html for more information.
ok: [virtual-machine-compute-rocky01-instance]

TASK [Display uptime] **********************************************************************************************************************************************************************************
ok: [virtual-machine-compute-rocky01-instance] => {
    "msg": "Uptime for virtual-machine-compute-rocky01-instance: 0:20:02"
}
ok: [virtual-machine-compute-rocky02-instance] => {
    "msg": "Uptime for virtual-machine-compute-rocky02-instance: 0:18:56"
}
ok: [virtual-machine-compute-rocky03-instance] => {
    "msg": "Uptime for virtual-machine-compute-rocky03-instance: 0:17:49"
}

TASK [Install nginx] ***********************************************************************************************************************************************************************************
ok: [virtual-machine-compute-rocky02-instance]
ok: [virtual-machine-compute-rocky03-instance]
ok: [virtual-machine-compute-rocky01-instance]

TASK [Start up nginx] **********************************************************************************************************************************************************************************
ok: [virtual-machine-compute-rocky03-instance]
ok: [virtual-machine-compute-rocky01-instance]
ok: [virtual-machine-compute-rocky02-instance]

TASK [Generate default page] ***************************************************************************************************************************************************************************
changed: [virtual-machine-compute-rocky03-instance]
changed: [virtual-machine-compute-rocky02-instance]
changed: [virtual-machine-compute-rocky01-instance]

PLAY RECAP *********************************************************************************************************************************************************************************************
virtual-machine-compute-rocky01-instance : ok=5    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
virtual-machine-compute-rocky02-instance : ok=5    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
virtual-machine-compute-rocky03-instance : ok=5    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

Looking good. I am almost done.

Final Steps - Load Balancing

I updated multiple-vm.yaml to create a service and route:

  - name: Create Service for the Rocky Linux deployment
    community.okd.k8s:
      definition:
        apiVersion: v1
        kind: Service
        metadata:
          name: rockylinux-built-with-ansible
          namespace: virtual-machine-compute
        spec:
          ports:
            - port: 80
              targetPort: 80
          selector:
            app: rockylinux

  - name: Expose the Rocky Linux service externally
    community.okd.openshift_route:
      service: rockylinux-built-with-ansible
      namespace: virtual-machine-compute
      termination: edge
      annotations:
        haproxy.router.openshift.io/balance: roundrobin
      port: 80
      tls:
        insecure_policy: redirect
      state: present
    register: route

Both service and route are fairly similar to what you see in a Kubernetes/OpenShift manifests, with some attributes moved around, so you will need to review the Ansible documents carefully against the Kubernetes documentation to avoid mis-placing certain attributes.

With that, the entire playbook looks like this:

---
- name: Create Virtual Machine Instances
  hosts: localhost
  connection: local
  vars:
    vm_list:
      - rocky01-instance
      - rocky02-instance
      - rocky03-instance
  tasks:
  - name: Create/Destroy VM instances
    loop: "{{ vm_list }}"
    kubevirt.core.kubevirt_vm:
      state: "{{ vm_state }}"
      wait: true
      name: "{{ item }}"
      namespace: virtual-machine-compute
      labels:
        app: rockylinux
      data_volume_templates:
        - metadata:
            name: "{{ item }}"
          spec:
            source:
              pvc:
                name: rockylinux-10-iscsi
                namespace: custom-vm-images
            pvc:
              accessModes:
                - ReadWriteMany
              resources:
                requests:
                  storage: 20Gi
              storageClassName: synology-iscsi-storage
              volumeMode: Block
      spec:
        evictionStrategy: LiveMigrate
        accessCredentials:
          - sshPublicKey:
              propagationMethod:
                noCloud: {}
              source:
                secret:
                  secretName: example-secret-public-key
        domain:
          cpu:
            cores: 1
            sockets: 1
            threads: 1
          memory:
            guest: 4Gi
          devices:
            disks:
              - disk:
                  bus: virtio
                name: rootdisk
              - disk:
                  bus: virtio
                name: cloudinitdisk
            interfaces:
              - masquerade: {}
                model: virtio
                name: default
              - bridge: {}
                model: virtio
                name: nic-white-chimpanzee-56
                state: up
        networks:
          - name: default
            pod: {}
          - multus:
              networkName: nad-efficient-grouse
            name: nic-white-chimpanzee-56
        volumes:
          - name: rootdisk
            dataVolume:
              name: "{{ item }}"
          - cloudInitNoCloud:
              networkData: |
                version: 2
                ethernets:
                  enp1s0:
                    dhcp4: true
                    # Higher metric means lower priority for the pod network
                    dhcp4-overrides:
                      route-metric: 200
                  enp2s0:
                    dhcp4: true
                    # Lower metric means this becomes the preferred default route
                    dhcp4-overrides:
                      route-metric: 50
                    # Explicit static route for your workstation network
                    routes:
                      - to: 192.168.1.0/24
                        via: 192.168.3.1
              userData: |-
                #cloud-config
                user: cloud-user
                chpasswd: { expire: False }
            name: cloudinitdisk

  - name: Create Service for the Rocky Linux deployment
    community.okd.k8s:
      definition:
        apiVersion: v1
        kind: Service
        metadata:
          name: rockylinux-built-with-ansible
          namespace: virtual-machine-compute
        spec:
          ports:
            - port: 80
              targetPort: 80
          selector:
            app: rockylinux

  - name: Expose the Rocky Linux service externally
    community.okd.openshift_route:
      service: rockylinux-built-with-ansible
      namespace: virtual-machine-compute
      termination: edge
      annotations:
        haproxy.router.openshift.io/balance: roundrobin
      port: 80
      tls:
        insecure_policy: redirect
      state: present
    register: route

I re-ran the playbook, which adds those components:

➜  kubevirt git:(main) ✗ uv run ansible-playbook multiple-vm.yaml -e vm_state=present

PLAY [Create Virtual Machine Instances] ****************************************************************************************************************************************************************

TASK [Gathering Facts] *********************************************************************************************************************************************************************************
ok: [localhost]

TASK [Create/Destroy VM instances] *********************************************************************************************************************************************************************
[WARNING]: No meaningful diff was generated, but the API may not be idempotent (only metadata.generation or metadata.resourceVersion were changed)
[WARNING]: Deprecation warnings can be disabled by setting `deprecation_warnings=False` in ansible.cfg.
[DEPRECATION WARNING]: Passing `warnings` to `exit_json` or `fail_json` is deprecated. This feature will be removed from ansible-core version 2.23. Use `AnsibleModule.warn` instead.
ok: [localhost] => (item=rocky01-instance)
ok: [localhost] => (item=rocky02-instance)
ok: [localhost] => (item=rocky03-instance)

TASK [Create Service for the rocky linux deployment] ***************************************************************************************************************************************************
changed: [localhost]

TASK [Expose the Rocky Linux service externally] *******************************************************************************************************************************************************
changed: [localhost]

PLAY RECAP *********************************************************************************************************************************************************************************************
localhost                  : ok=4    changed=2    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

I then ran curl against the route to verify connectivity and confirm all the instances are up and running, looping multiple times to confirm that traffic is going to each node:

➜  kubevirt git:(main) ✗ for count in `seq 1 33`
for> do
for> curl https://rockylinux-built-with-ansible-virtual-machine-compute.apps.okd.example.com
for> echo
for> done
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky01-instance: 0:20:02
Uptime for virtual-machine-compute-rocky03-instance: 0:17:49
Uptime for virtual-machine-compute-rocky02-instance: 0:18:56
➜  kubevirt git:(main) ✗

Resources