ci: add upgrade

    - Add upgrade test on CI. The detail steps as below

      1. Install latest version
      2. Add disk
      3. Upgrade to the develping image
      4. Run integration test

Signed-off-by: Vicente Cheng <vicente.cheng@suse.com>

.github/workflows/basic-ci.yaml

@@ -51,12 +51,22 @@ jobs:
           cp ../ci/scripts/deploy_ndm.sh ./deploy_ndm.sh
           ./deploy_ndm.sh
           popd
-      - name: "Basic Test"
-        id: basic-test
+      - name: "Add disk"
         run: |
           pushd ndm-vagrant-rancherd
           ./scripts/attach-disk.sh node1
           sleep 30
+          popd
+      - name: "Upgrade NDM"
+        run: |
+          pushd ndm-vagrant-rancherd
+          cp ../ci/scripts/upgrade_ndm.sh ./upgrade_ndm.sh
+          ./upgrade_ndm.sh
+          popd
+      - name: "Run Basic Test"
+        id: basic-test
+        run: |
+          pushd ndm-vagrant-rancherd
           vagrant ssh-config node1 > ../ssh-config
           cp kubeconfig ../kubeconfig
           popd

ci/scripts/deploy_ndm.sh

@@ -83,24 +83,14 @@ echo "cluster nodes: $cluster_nodes"
 ensure_longhorn_ready
 
 pushd $TOP_DIR
-# cleanup first
-rm -rf harvester-node-disk-manager
 
-cp -r ../deploy/charts/harvester-node-disk-manager harvester-node-disk-manager
-cp ../ci/charts/ndm-override.yaml ndm-override.yaml
+cat >> ndm-override.yaml << 'EOF'
+autoProvisionFilter: [/dev/sd*]
+EOF
 
-$HELM install -f $TOP_DIR/ndm-override.yaml harvester-node-disk-manager harvester-node-disk-manager/ --create-namespace -n harvester-system
+$HELM pull harvester-node-disk-manager --repo https://charts.harvesterhci.io --untar
+$HELM install -f $TOP_DIR/ndm-override.yaml harvester-node-disk-manager ./harvester-node-disk-manager --create-namespace -n harvester-system
 
 wait_ndm_ready
-# check image
-pod_name=$(kubectl get pods -n harvester-system |grep ^harvester-node-disk-manager|head -n1 |awk '{print $1}')
-container_img=$(kubectl get pods ${pod_name} -n harvester-system -o yaml |yq -e .spec.containers[0].image |tr ":" \n)
-yaml_img=$(yq -e .image.repository ndm-override.yaml)
-if grep -q ${yaml_img} <<< ${container_img}; then
-  echo "Image is equal: ${yaml_img}"
-else
-  echo "Image is non-equal, container: ${container_img}, yaml file: ${yaml_img}"
-  exit 1
-fi
 echo "harvester-node-disk-manager is ready"
 popd

ci/scripts/upgrade_ndm.sh

@@ -0,0 +1,73 @@
+#!/bin/bash -e
+
+TOP_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )/" &> /dev/null && pwd )"
+
+ensure_command() {
+  local cmd=$1
+  if ! which $cmd &> /dev/null; then
+    echo 1
+    return
+  fi
+  echo 0
+}
+
+wait_ndm_ready() {
+  while [ true ]; do
+    running_num=$(kubectl get ds harvester-node-disk-manager -n harvester-system -o 'jsonpath={.status.numberReady}')
+    if [[ $running_num -eq ${cluster_nodes} ]]; then
+      echo "harvester-node-disk-manager pods are ready!"
+      break
+    fi
+    echo "harvester-node-disk-manager pods are not ready (${running_num}/${cluster_nodes}), sleep 10 seconds."
+    sleep 10
+  done
+}
+
+if [ ! -f $TOP_DIR/kubeconfig ]; then
+  echo "kubeconfig does not exist. Please create cluster first."
+  echo "Maybe try new_cluster.sh"
+  exit 1
+fi
+echo $TOP_DIR/kubeconfig
+export KUBECONFIG=$TOP_DIR/kubeconfig
+
+if [[ $(ensure_command helm) -eq 1 ]]; then
+  echo "no helm, try to curl..."
+  curl -O https://get.helm.sh/helm-v3.9.4-linux-amd64.tar.gz
+  tar -zxvf helm-v3.9.4-linux-amd64.tar.gz
+  HELM=$TOP_DIR/linux-amd64/helm
+  $HELM version
+else
+  echo "Get helm, version info as below"
+  HELM=$(which helm)
+  $HELM version
+fi
+
+cluster_nodes=$(yq -e e '.cluster_size' $TOP_DIR/settings.yaml)
+echo "cluster nodes: $cluster_nodes"
+
+pushd $TOP_DIR
+# cleanup first
+rm -rf harvester-node-disk-manager*
+rm -rf ndm-override.yaml
+
+cp -r ../deploy/charts/harvester-node-disk-manager harvester-node-disk-manager
+cp ../ci/charts/ndm-override.yaml ndm-override.yaml
+
+$HELM upgrade -f $TOP_DIR/ndm-override.yaml harvester-node-disk-manager harvester-node-disk-manager/ -n harvester-system
+
+sleep 10 # wait 10 seconds for ndm start to respwan pods
+
+wait_ndm_ready
+# check image
+pod_name=$(kubectl get pods -n harvester-system |grep ^harvester-node-disk-manager|head -n1 |awk '{print $1}')
+container_img=$(kubectl get pods ${pod_name} -n harvester-system -o yaml |yq -e .spec.containers[0].image |tr ":" \n)
+yaml_img=$(yq -e .image.repository ndm-override.yaml)
+if grep -q ${yaml_img} <<< ${container_img}; then
+  echo "Image is equal: ${yaml_img}"
+else
+  echo "Image is non-equal, container: ${container_img}, yaml file: ${yaml_img}"
+  exit 1
+fi
+echo "harvester-node-disk-manager upgrade successfully!"
+popd

deploy/charts/harvester-node-disk-manager/Chart.yaml

@@ -15,13 +15,13 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.4.8
+version: 0.4.11
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to
 # follow Semantic Versioning. They should reflect the version the application is using.
 # It is recommended to use it with quotes.
-appVersion: v0.4.8
+appVersion: 0.4.11
 
 maintainers:
   - name: harvester

deploy/charts/harvester-node-disk-manager/templates/crds/harvesterhci.io_blockdevices.yaml

@@ -32,8 +32,8 @@ spec:
     - jsonPath: .spec.nodeName
       name: NodeName
       type: string
-    - jsonPath: .status.state
-      name: Status
+    - jsonPath: .status.provisionPhase
+      name: ProvisionPhase
       type: string
     - jsonPath: .metadata.creationTimestamp
       name: Age
@@ -66,13 +66,18 @@ spec:
                       overwrite the existing one
                     type: boolean
                   mountPoint:
-                    description: a string with the partition's mount point, or ""
-                      if no mount point was discovered
+                    description: 'DEPRECATED: no longer use and has no effect. a string
+                      with the partition''s mount point, or "" if no mount point was
+                      discovered'
                     type: string
                   provisioned:
                     description: a bool indicating whether the filesystem can be provisioned
                       as a disk for the node to store data.
                     type: boolean
+                  repaired:
+                    description: a bool indicating whether the filesystem is manually
+                      repaired of not
+                    type: boolean
                 required:
                 - mountPoint
                 type: object
@@ -223,6 +228,10 @@ spec:
                           when user operate device formatting through the CRD controller
                         format: date-time
                         type: string
+                      corrupted:
+                        description: indicating whether the filesystem is corrupted
+                          or not
+                        type: boolean
                       isReadOnly:
                         description: a bool indicating the partition is read-only
                         type: boolean