CI: convert vagrant jobs to kubevirt

Vagrant jobs needs a big cache which makes them slow / sometimes stuck
completely. Using the kubevirt provisionning playbook is now
significantly faster, so do just that.

Having only one provisionner in CI will also allows us to remove some of
the custom runners executors we use for vagrant, and more generally
reduce the CI maintenance.

Our kubevirt CI platform does not support ivp6 yet, so we keep the
relevant jobs in vagrant, but we'll migrate them as well as soon as
possible.

.gitlab-ci/kubevirt.yml

@@ -42,12 +42,17 @@ pr:
           - debian11-calico-collection
           - debian11-macvlan
           - debian12-cilium
+          - fedora39-kube-router
+            # FIXME: this test if broken (perma-failing)
           - openeuler24-calico
           - opensuse15-6-calico
           - rockylinux8-calico
           - rockylinux9-cilium
           - ubuntu20-calico-all-in-one-hardening
           - ubuntu20-cilium-sep
+          - ubuntu20-flannel-collection
+          - ubuntu20-kube-router-sep
+          - ubuntu20-kube-router-svc-proxy
           - ubuntu22-calico-all-in-one
           - ubuntu22-calico-all-in-one-upgrade
           - ubuntu24-calico-etcd-datastore
@@ -120,6 +125,7 @@ pr_extended:
           - opensuse15-6-docker-cilium
           - rockylinux9-calico
           - ubuntu20-calico-etcd-kubeadm
+          - ubuntu20-flannel
           - ubuntu22-all-in-one-docker
           - ubuntu24-all-in-one-docker
           - ubuntu24-calico-all-in-one

.gitlab-ci/vagrant.yml

@@ -1,13 +1,13 @@
 ---
-.vagrant:
-  extends: .testcases
+vagrant:
+  extends: .job-moderated
   needs:
     - ci-not-authorized
   variables:
     CI_PLATFORM: "vagrant"
     SSH_USER: "vagrant"
     VAGRANT_DEFAULT_PROVIDER: "libvirt"
-    KUBESPRAY_VAGRANT_CONFIG: tests/files/${CI_JOB_NAME}.rb
+    KUBESPRAY_VAGRANT_CONFIG: tests/files/${TESTCASE}.rb
     DOCKER_NAME: vagrant
     VAGRANT_ANSIBLE_TAGS: facts
     VAGRANT_HOME: "$CI_PROJECT_DIR/.vagrant.d"
@@ -35,47 +35,12 @@
       - .vagrant.d/boxes
       - .cache/pip
     policy: pull-push # TODO: change to "pull" when not on main
-
-vagrant_ubuntu24-calico-dual-stack:
   stage: deploy-extended
-  extends: .vagrant
   rules:
     - if: $PR_LABELS =~ /.*(ci-extended|ci-full).*/
       when: on_success
-  allow_failure: false
-
-vagrant_ubuntu24-calico-ipv6only-stack:
-  stage: deploy-extended
-  extends: .vagrant
-  rules:
-    - if: $PR_LABELS =~ /.*(ci-extended|ci-full).*/
-      when: on_success
-  allow_failure: false
-
-vagrant_ubuntu20-flannel:
-  stage: deploy-part1
-  extends: .vagrant
-  when: on_success
-  allow_failure: false
-
-vagrant_ubuntu20-flannel-collection:
-  stage: deploy-extended
-  extends: .vagrant
-  when: manual
-
-vagrant_ubuntu20-kube-router-sep:
-  stage: deploy-extended
-  extends: .vagrant
-  when: manual
-
-# Service proxy test fails connectivity testing
-vagrant_ubuntu20-kube-router-svc-proxy:
-  stage: deploy-extended
-  extends: .vagrant
-  when: manual
-
-vagrant_fedora39-kube-router:
-  stage: deploy-extended
-  extends: .vagrant
-  when: manual
-# FIXME: this test if broken (perma-failing)
+  parallel:
+    matrix:
+      - TESTCASE:
+          - ubuntu24-calico-dual-stack
+          - ubuntu24-calico-ipv6only-stack

tests/files/fedora39-kube-router.yml

@@ -0,0 +1,7 @@
+---
+cloud_image: fedora-39
+cluster_layout:
+  - node_groups: ['kube_control_plane', 'etcd', 'kube_node']
+  - node_groups: ['kube_node']
+
+kube_network_plugin: "kube-router"

tests/files/ubuntu20-flannel-collection.yml

@@ -0,0 +1,8 @@
+---
+cloud_image: ubuntu-2004
+cluster_layout:
+  - node_groups: ['kube_control_plane', 'etcd', 'kube_node']
+  - node_groups: ['kube_control_plane', 'etcd', 'kube_node']
+  - node_groups: ['etcd', 'kube_node']
+
+kube_network_plugin: flannel

tests/files/ubuntu20-flannel.yml

@@ -0,0 +1 @@
+ubuntu20-flannel-collection.yml

tests/files/ubuntu20-kube-router-sep.yml

@@ -0,0 +1,7 @@
+---
+cloud_image: ubuntu-2004
+cluster_layout:
+  - node_groups: ['kube_control_plane', 'etcd', 'kube_node']
+  - node_groups: ['kube_node']
+
+kube_network_plugin: "kube-router"

tests/files/ubuntu20-kube-router-svc-proxy.yml

@@ -0,0 +1,10 @@
+---
+cloud_image: ubuntu-2004
+cluster_layout:
+  - node_groups: ['kube_control_plane', 'etcd', 'kube_node']
+  - node_groups: ['kube_control_plane', 'etcd', 'kube_node']
+  - node_groups: ['etcd', 'kube_node']
+
+kube_network_plugin: "kube-router"
+
+kube_router_run_service_proxy: true

tests/files/vagrant_fedora39-kube-router.rb

@@ -1,15 +0,0 @@
-$num_instances = 2
-$vm_memory ||= 2048
-$os = "fedora39"
-
-$control_plane_instances = 1
-$etcd_instances = 1
-
-# For CI we are not worried about data persistence across reboot
-$libvirt_volume_cache = "unsafe"
-
-# Checking for box update can trigger API rate limiting
-# https://www.vagrantup.com/docs/vagrant-cloud/request-limits.html
-$box_check_update = false
-
-$network_plugin = "kube-router"

tests/files/vagrant_fedora39-kube-router.yml

@@ -1,6 +0,0 @@
----
-# Instance settings
-cloud_image: fedora-39
-
-# Kubespray settings
-kube_network_plugin: kube-router

tests/files/vagrant_ubuntu20-flannel-collection.rb

@@ -1,9 +0,0 @@
-$os = "ubuntu2004"
-
-# For CI we are not worries about data persistence across reboot
-$libvirt_volume_cache = "unsafe"
-
-# Checking for box update can trigger API rate limiting
-# https://www.vagrantup.com/docs/vagrant-cloud/request-limits.html
-$box_check_update = false
-$vm_cpus = 2

tests/files/vagrant_ubuntu20-flannel-collection.yml

@@ -1,3 +0,0 @@
----
-# Kubespray settings
-kube_network_plugin: flannel

tests/files/vagrant_ubuntu20-flannel.rb

@@ -1,8 +0,0 @@
-$os = "ubuntu2004"
-
-# For CI we are not worries about data persistence across reboot
-$libvirt_volume_cache = "unsafe"
-# Checking for box update can trigger API rate limiting
-# https://www.vagrantup.com/docs/vagrant-cloud/request-limits.html
-$box_check_update = false
-$vm_cpus = 2

tests/files/vagrant_ubuntu20-flannel.yml

@@ -1,4 +0,0 @@
----
-# Kubespray settings
-kube_network_plugin: flannel
-ansible_ssh_private_key: .vagrant.d/insecure_private_key

tests/files/vagrant_ubuntu20-kube-router-sep.rb

@@ -1,15 +0,0 @@
-$num_instances = 2
-$vm_memory ||= 2048
-$os = "ubuntu2004"
-
-$control_plane_instances = 1
-$etcd_instances = 1
-
-# For CI we are not worried about data persistence across reboot
-$libvirt_volume_cache = "unsafe"
-
-# Checking for box update can trigger API rate limiting
-# https://www.vagrantup.com/docs/vagrant-cloud/request-limits.html
-$box_check_update = false
-
-$network_plugin = "kube-router"

tests/files/vagrant_ubuntu20-kube-router-sep.yml

@@ -1,8 +0,0 @@
----
-# Instance settings
-cloud_image: ubuntu-2004
-mode: separate
-
-# Kubespray settings
-bootstrap_os: ubuntu
-kube_network_plugin: kube-router

tests/files/vagrant_ubuntu20-kube-router-svc-proxy.rb

@@ -1,10 +0,0 @@
-$os = "ubuntu2004"
-
-# For CI we are not worried about data persistence across reboot
-$libvirt_volume_cache = "unsafe"
-
-# Checking for box update can trigger API rate limiting
-# https://www.vagrantup.com/docs/vagrant-cloud/request-limits.html
-$box_check_update = false
-
-$network_plugin = "kube-router"

tests/files/vagrant_ubuntu20-kube-router-svc-proxy.yml

@@ -1,10 +0,0 @@
----
-# Instance settings
-cloud_image: ubuntu-2004
-mode: separate
-
-# Kubespray settings
-bootstrap_os: ubuntu
-kube_network_plugin: kube-router
-
-kube_router_run_service_proxy: true