Merge pull request #12037 from VannTen/ci/convert_vagrant_to_kubevirt_2

CI: convert remaining vagrant jobs (except IPv6) to kubevirt + cleanups

tests/Makefile

@@ -8,7 +8,7 @@ $(ANSIBLE_INVENTORY):
 	mkdir $@
 
 create-packet: | $(ANSIBLE_INVENTORY)
-	ansible-playbook cloud_playbooks/create-packet.yml -c local \
+	ansible-playbook cloud_playbooks/create-kubevirt.yml -c local \
 		-e @"files/${CI_JOB_NAME}.yml"
 
 delete-packet: ;

tests/cloud_playbooks/roles/packet-ci/tasks/main.yml

@@ -1,8 +1,4 @@
 ---
-- name: Include custom vars for ci job
-  include_vars: "../files/{{ lookup('ansible.builtin.env', 'CI_JOB_NAME') }}.yml"
-  when: molecule_yml is not defined
-
 - name: Generate SSH keypair
   community.crypto.openssh_keypair:
     size: 2048

tests/files/almalinux9-calico-remove-node

@@ -0,0 +1,2 @@
+REMOVE_NODE_CHECK=true
+REMOVE_NODE_NAME=instance-3

tests/files/debian11-calico-upgrade

@@ -0,0 +1 @@
+UPGRADE_TEST=graceful

tests/files/debian11-calico-upgrade-once

@@ -0,0 +1 @@
+UPGRADE_TEST=graceful

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/fedora40-docker.calico

@@ -0,0 +1 @@
+RESET_CHECK=true

tests/files/rockylinux9-cilium

@@ -0,0 +1 @@
+RESET_CHECK=true

tests/files/ubuntu20-calico-all-in-one

@@ -0,0 +1 @@
+RESET_CHECK=true

tests/files/ubuntu20-calico-etcd-kubeadm-upgrade-ha

@@ -0,0 +1 @@
+UPGRADE_TEST=basic

tests/files/ubuntu20-calico-ha-recover

@@ -0,0 +1,2 @@
+RECOVER_CONTROL_PLANE_TEST=true
+RECOVER_CONTROL_PLANE_TEST_GROUPS="etcd[2:]:kube_control_plane[1:]"

tests/files/ubuntu20-calico-ha-recover-noquorum

@@ -0,0 +1,2 @@
+RECOVER_CONTROL_PLANE_TEST=true
+RECOVER_CONTROL_PLANE_TEST_GROUPS="etcd[1:]:kube_control_plane[1:]"

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/ubuntu22-calico-all-in-one-upgrade

@@ -0,0 +1 @@
+UPGRADE_TEST=graceful

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

tests/scripts/create-tf.sh

@@ -1,5 +0,0 @@
-#!/bin/bash
-set -euxo pipefail
-
-cd ..
-terraform -chdir="contrib/terraform/$PROVIDER" apply -auto-approve -parallelism=1

tests/scripts/delete-tf.sh

@@ -1,5 +0,0 @@
-#!/bin/bash
-set -euxo pipefail
-
-cd ..
-terraform -chdir="contrib/terraform/$PROVIDER" destroy -auto-approve

tests/scripts/md-table/main.py

@@ -90,7 +90,7 @@ for f in files:
 
     container_manager = y.get('container_manager', 'containerd')
     network_plugin = y.get('kube_network_plugin', 'calico')
-    x = re.match(r"^[a-z-]+_([a-z0-9]+).*", f.name)
-    operating_system = x.group(1)
+    x = re.match(r"^([a-z-]+_)?([a-z0-9]+).*", f.name)
+    operating_system = x.group(2)
     data.set(container_manager=container_manager, network_plugin=network_plugin, os=operating_system)
 print(data.jinja(), file=open(args.output, 'w'))

tests/scripts/testcases_cleanup.sh

@@ -1,4 +0,0 @@
-#!/bin/bash
-set -euxo pipefail
-
-make -C tests delete-${CI_PLATFORM} -s

tests/scripts/testcases_prepare.sh

@@ -1,7 +0,0 @@
-#!/bin/bash
-set -euxo pipefail
-
-mkdir -p /.ssh
-mkdir -p cluster-dump
-mkdir -p $HOME/.ssh
-ansible-playbook --version

tests/scripts/testcases_run.sh

@@ -1,20 +1,17 @@
 #!/bin/bash
 set -euxo pipefail
 
-echo "CI_JOB_NAME is $CI_JOB_NAME"
-
-if [[ "$CI_JOB_NAME" =~ "upgrade" ]]; then
-  if [ "${UPGRADE_TEST}" == "false" ]; then
-    echo "Job name contains 'upgrade', but UPGRADE_TEST='false'"
-    exit 1
-  fi
+if [[ -v TESTCASE ]]; then
+    TESTCASE_FILE=files/${TESTCASE}.yml
 else
-  if [ "${UPGRADE_TEST}" != "false" ]; then
-    echo "UPGRADE_TEST!='false', but job names does not contain 'upgrade'"
-    exit 1
-  fi
+    TESTCASE_FILE=common_vars.yml
+    TESTCASE=default
 fi
 
+echo "TESTCASE is $TESTCASE"
+
+source tests/files/$TESTCASE || true
+
 # Check out latest tag if testing upgrade
 if [ "${UPGRADE_TEST}" != "false" ]; then
   git fetch --all && git checkout $(git describe --tags --abbrev=0)
@@ -26,12 +23,10 @@ fi
 export ANSIBLE_BECOME=true
 export ANSIBLE_BECOME_USER=root
 
-make -C tests create-${CI_PLATFORM} -s
-
 # Test collection build and install by installing our collection, emptying our repository, adding
 # cluster.yml, reset.yml, and remote-node.yml files that simply point to our collection's playbooks, and then
 # running the same tests as before
-if [[ "${CI_JOB_NAME}" =~ "collection" ]]; then
+if [[ "${TESTCASE}" =~ "collection" ]]; then
   # Build and install collection
   ansible-galaxy collection build
   ansible-galaxy collection install kubernetes_sigs-kubespray-$(grep "^version:" galaxy.yml | awk '{print $2}').tar.gz
@@ -66,10 +61,9 @@ fi
 run_playbook () {
 playbook=$1
 shift
-# We can set --limit here and still pass it as supplemental args because `--limit`  is a 'last one wins' option
 ansible-playbook \
     -e @tests/common_vars.yml \
-    -e @tests/files/${CI_JOB_NAME}.yml \
+    -e @tests/${TESTCASE_FILE} \
     -e local_release_dir=${PWD}/downloads \
     "$@" \
     ${playbook}
@@ -111,10 +105,10 @@ run_playbook tests/testcases/015_check-nodes-ready.yml
 
 ## Test that all nodes are Ready
 
-if [[ ! ( "$CI_JOB_NAME" =~ "macvlan" ) ]]; then
+if [[ ! ( "$TESTCASE" =~ "macvlan" ) ]]; then
     run_playbook tests/testcases/020_check-pods-running.yml
     run_playbook tests/testcases/030_check-network.yml
-    if [[ ! ( "$CI_JOB_NAME" =~ "hardening" ) ]]; then
+    if [[ ! ( "$TESTCASE" =~ "hardening" ) ]]; then
       # TODO: We need to remove this condition by finding alternative container
       # image instead of netchecker which doesn't work at hardening environments.
       run_playbook tests/testcases/040_check-network-adv.yml