testing: add workflows for testing kata-deploy

1. AKS based action updated to be run from either packaging or remote
repository. We will only clone kata-deploy for yaml/scripts/tests if we
are running the action outside of the packaging repo. If in packaging,
the bits are already included. Misc. cleanup as well.

2. Workflow introduced which leverages the updated AKS action. This will
allow testing of packaging changes to kata-deploy.

The workflow itself uses the following github action: xt0rted/slash-command-action

The workflow will create a kata-deploy container image based off of the latest
release, utilizing the latest released Kata artifacts off of master. It
will then use the AKS kata-deploy GitHub action.

Users with admin access on the repo can trigger this test by:
/test kata-deploy

Fixes: #845

Signed-off-by: Eric Ernst <eric.ernst@intel.com>

.github/workflows/kata-deploy-test.yaml

@@ -0,0 +1,54 @@
+on: issue_comment
+name: test-kata-deploy
+jobs:
+  check_comments:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check for Command
+        id: command
+        uses: kata-containers/slash-command-action@v1
+        with:
+          repo-token: ${{ secrets.GITHUB_TOKEN }}
+          command: "test"
+          reaction: "true"
+          reaction-type: "eyes"
+          allow-edits: "false"
+          permission-level: admin
+      - name: verify command arg is kata-deploy
+        run: |
+           echo "The command was '${{ steps.command.outputs.command-name }}' with arguments '${{ steps.command.outputs.command-arguments }}'"
+           [[ ${{ steps.command.outputs.command-arguments}} == "kata-deploy" ]]
+  create-and-test-container:
+    needs: check_comments
+    runs-on: ubuntu-latest
+    steps:
+      - name: get-PR-ref
+        id: get-PR-ref
+        run: |
+            ref=$(cat $GITHUB_EVENT_PATH | jq -r '.issue.pull_request.url' | sed  's#^.*\/pulls#refs\/pull#' | sed 's#$#\/merge#')
+            echo "reference for PR: " ${ref}
+            echo "##[set-output name=pr-ref;]${ref}"
+      - uses: actions/checkout@v2-beta
+        with:
+          ref: ${{ steps.get-PR-ref.outputs.pr-ref }}
+      - name: build-container-image
+        id: build-container-image
+        run: |
+            PR_SHA=$(git log --format=format:%H -n1)
+            VERSION=$(curl https://raw.githubusercontent.com/kata-containers/runtime/master/VERSION)
+            ARTIFACT_URL="https://github.com/kata-containers/runtime/releases/download/${VERSION}/kata-static-${VERSION}-x86_64.tar.xz"
+            wget "${ARTIFACT_URL}" -O ./kata-deploy/kata-static.tar.xz
+            docker build --build-arg KATA_ARTIFACTS=kata-static.tar.xz -t katadocker/kata-deploy-ci:${PR_SHA} ./kata-deploy
+            docker login -u ${{ secrets.DOCKER_USERNAME }} -p ${{ secrets.DOCKER_PASSWORD }}
+            docker push katadocker/kata-deploy-ci:$PR_SHA
+            echo "##[set-output name=pr-sha;]${PR_SHA}"
+      - name: test-kata-deploy-ci-in-aks
+        uses: ./kata-deploy/action
+        with:
+          packaging-sha: ${{ steps.build-container-image.outputs.pr-sha }}
+        env:
+          PKG_SHA: ${{ steps.build-container-image.outputs.pr-sha }}
+          AZ_APPID: ${{ secrets.AZ_APPID }}
+          AZ_PASSWORD: ${{ secrets.AZ_PASSWORD }}
+          AZ_SUBSCRIPTION_ID: ${{ secrets.AZ_SUBSCRIPTION_ID }}
+          AZ_TENANT_ID: ${{ secrets.AZ_TENANT_ID }}

kata-deploy/action/action.yml

@@ -1,9 +1,9 @@
 # action.yml
-name: 'kata-deploy'
+name: 'kata-deploy-aks'
 description: 'test Kata container image in AKS'
 inputs:
   packaging-sha:
-    description: 'SHA we are using for pulling packaing manifests'
+    description: 'SHA we are using for pulling packaging manifests'
     required: true
     default: ''
 runs:

kata-deploy/action/setup-aks.sh

@@ -23,9 +23,7 @@ function die() {
 function destroy_aks() {
     set +x
 
-    export KUBECONFIG="_output/$DNS_PREFIX/kubeconfig/kubeconfig.$LOCATION.json"
-    kubectl describe ds -n kube-system kata-deploy || true
-    kubectl describe ds -n kube-system kata-cleanup || true
+    export KUBECONFIG="$PWD/_output/$DNS_PREFIX/kubeconfig/kubeconfig.$LOCATION.json"
 
     az login --service-principal -u "$AZ_APPID" -p "$AZ_PASSWORD" --tenant "$AZ_TENANT_ID"
     az group delete --name "$DNS_PREFIX" --yes --no-wait

kata-deploy/action/test-kata.sh

@@ -17,13 +17,13 @@ function die() {
 
 function waitForProcess() {
     wait_time="$1"
-    sleep_time="$2"
-    cmd="$3"
+    cmd="$2"
+    sleep_time=5
+    echo "waiting for process $cmd"
     while [ "$wait_time" -gt 0 ]; do
         if eval "$cmd"; then
             return 0
         else
-            echo "waiting"
             sleep "$sleep_time"
             wait_time=$((wait_time-sleep_time))
         fi
@@ -35,16 +35,16 @@ function waitForProcess() {
 # timeout expires
 function waitForLabelRemoval() {
     wait_time="$1"
-    sleep_time="$2"
+    sleep_time=5
 
+    echo "waiting for kata-runtime label to be removed"
     while [[ "$wait_time" -gt 0 ]]; do
         # if a node is found which matches node-select, the output will include a column for node name,
         # NAME. Let's look for that 
-        if [[ -z $(kubectl get nodes --selector katacontainers.io/kata-runtime | grep NAME) ]]
+        if [[ -z $(kubectl get nodes --selector katacontainers.io/kata-runtime 2>&1 | grep NAME) ]]
         then
             return 0
         else
-            echo "waiting for kata-runtime label to be removed"
             sleep "$sleep_time"
             wait_time=$((wait_time-sleep_time))
         fi
@@ -56,10 +56,8 @@ function waitForLabelRemoval() {
     return 1
 }
 
-
 function run_test() {
-    PKG_SHA=$1
-    YAMLPATH="https://raw.githubusercontent.com/kata-containers/packaging/$PKG_SHA/kata-deploy"
+    YAMLPATH="./kata-deploy"
     echo "verify connectivity with a pod using Kata"
 
     deployment=""
@@ -67,7 +65,6 @@ function run_test() {
     busybox_image="busybox"
     cmd="kubectl get pods | grep $busybox_pod | grep Completed"
     wait_time=120
-    sleep_time=3
 
     configurations=("nginx-deployment-qemu" "nginx-deployment-qemu-virtiofs")
     for deployment in "${configurations[@]}"; do
@@ -83,7 +80,7 @@ function run_test() {
 
       # test pod connectivity:
       kubectl run $busybox_pod --restart=Never --image="$busybox_image" -- wget --timeout=5 "$deployment"
-      waitForProcess "$wait_time" "$sleep_time" "$cmd"
+      waitForProcess "$wait_time" "$cmd"
       kubectl logs "$busybox_pod" | grep "index.html"
       kubectl describe pod "$busybox_pod"
 
@@ -99,12 +96,19 @@ function test_kata() {
     set -x
 
     [[ -z "$PKG_SHA" ]] && die "no PKG_SHA provided"
-    echo "$PKG_SHA"
 
-    #kubectl all the things
-    kubectl get pods,nodes --all-namespaces
+    # This action could be called in two contexts:
+    #  1. Packaging workflows: testing in packaging repository, where we assume yaml/packaging
+    #   bits under test are already part of teh action workspace.
+    #  2. From kata-containers: when creating a release, the appropriate packaging repository is
+    #   not yet part of the workspace, and we will need to clone
+    if [[ ! -d ./kata-deploy ]]; then
+        git clone https://github.com/kata-containers/packaging packaging
+        cd packaging
+        git checkout $PKG_SHA
+    fi
 
-    YAMLPATH="https://raw.githubusercontent.com/kata-containers/packaging/$PKG_SHA/kata-deploy"
+    YAMLPATH="./kata-deploy"
 
     kubectl apply -f "$YAMLPATH/kata-rbac/base/kata-rbac.yaml"
 
@@ -114,17 +118,14 @@ function test_kata() {
 
     kubectl get runtimeclasses
 
-    curl -LO "$YAMLPATH/kata-deploy/base/kata-deploy.yaml"
-    curl -LO "$YAMLPATH/kata-cleanup/base/kata-cleanup.yaml"
-
     # update deployment daemonset to utilize the container under test:
-    sed -i "s#katadocker/kata-deploy#katadocker/kata-deploy-ci:${PKG_SHA}#g" kata-deploy.yaml
-    sed -i "s#katadocker/kata-deploy#katadocker/kata-deploy-ci:${PKG_SHA}#g" kata-cleanup.yaml
+    sed -i "s#katadocker/kata-deploy#katadocker/kata-deploy-ci:${PKG_SHA}#g" $YAMLPATH/kata-deploy/base/kata-deploy.yaml
+    sed -i "s#katadocker/kata-deploy#katadocker/kata-deploy-ci:${PKG_SHA}#g" $YAMLPATH/kata-cleanup/base/kata-cleanup.yaml
 
-    cat kata-deploy.yaml
+    cat $YAMLPATH/kata-deploy/base/kata-deploy.yaml
 
     # deploy kata:
-    kubectl apply -f kata-deploy.yaml
+    kubectl apply -f $YAMLPATH/kata-deploy/base/kata-deploy.yaml
 
     # in case the control plane is slow, give it a few seconds to accept the yaml, otherwise
     # our 'wait' for deployment status will fail to find the deployment at all. If it can't persist
@@ -137,29 +138,25 @@ function test_kata() {
     # show running pods, and labels of nodes
     kubectl get pods,nodes --all-namespaces --show-labels
 
-    run_test $PKG_SHA
+    run_test
 
     kubectl get pods,nodes --show-labels
 
     # Remove Kata
-    kubectl delete -f kata-deploy.yaml
+    kubectl delete -f $YAMLPATH/kata-deploy/base/kata-deploy.yaml
     kubectl -n kube-system wait --timeout=10m --for=delete -l name=kata-deploy pod
 
     kubectl get pods,nodes --show-labels
 
-    kubectl apply -f kata-cleanup.yaml
+    kubectl apply -f $YAMLPATH/kata-cleanup/base/kata-cleanup.yaml
 
     # The cleanup daemonset will run a single time, since it will clear the node-label. Thus, its difficult to
     # check the daemonset's status for completion. instead, let's wait until the kata-runtime labels are removed
     # from all of the worker nodes. If this doesn't happen after 2 minutes, let's fail
-    timeout=20
-    sleeptime=6
-    waitForLabelRemoval $timeout $sleeptime
+    timeout=120
+    waitForLabelRemoval $timeout
 
-    kubectl delete -f kata-cleanup.yaml
-
-    rm kata-cleanup.yaml
-    rm kata-deploy.yaml
+    kubectl delete -f $YAMLPATH/kata-cleanup/base/kata-cleanup.yaml
 
     set +x
 }

kata-deploy/scripts/kata-deploy.sh

@@ -32,18 +32,19 @@ function print_usage() {
 }
 
 function get_container_runtime() {
-	local runtime="$(kubectl describe node $NODE_NAME)"
+
+	local runtime=$(kubectl get node $NODE_NAME -o jsonpath='{.status.nodeInfo.containerRuntimeVersion}' | awk -F '[:]' '{print $1}')
 	if [ "$?" -ne 0 ]; then
                 die "invalid node name"
 	fi
-	if echo "$runtime" | grep -qE 'Container Runtime Version.*containerd.*-k3s'; then
+	if echo "$runtime" | grep -qE 'containerd.*-k3s'; then
 		if systemctl is-active --quiet k3s-agent; then
 			echo "k3s-agent"
 		else
 			echo "k3s"
 		fi
 	else
-		echo "$runtime" | awk -F'[:]' '/Container Runtime Version/ {print $2}' | tr -d ' '
+		echo "$runtime"
 	fi
 }