Merge pull request #110 from senthilrch/develop

Merge commits for v0.8.2 release
senthilrch · Sep 1, 2021 · 088024e · 088024e
2 parents b7e37f4 + 304712a
commit 088024e
Show file tree

Hide file tree

Showing 13 changed files with 126 additions and 82 deletions.
diff --git a/Makefile b/Makefile
@@ -43,7 +43,7 @@ ifndef OPERATOR_IMAGE_REPO
 endif
 
 ifndef RELEASE_VERSION
-  RELEASE_VERSION=v0.8.1
+  RELEASE_VERSION=v0.8.2
 endif
 
 ifndef DOCKER_VERSION
@@ -82,10 +82,6 @@ ifndef PROGRESS
   PROGRESS=auto
 endif
 
-ifndef OPERATOR_NAMESPACE
-  OPERATOR_NAMESPACE=kubefledged-operator
-endif
-
 ifndef KUBEFLEDGED_NAMESPACE
   KUBEFLEDGED_NAMESPACE=kube-fledged
 endif
@@ -206,22 +202,20 @@ deploy-using-yaml:
 	kubectl rollout status deployment kubefledged-controller -n kube-fledged --watch
 
 deploy-using-operator:
-	# Create the namespaces for operator and kubefledged
-	-kubectl create namespace ${OPERATOR_NAMESPACE}
+	# Create the namespace
 	-kubectl create namespace ${KUBEFLEDGED_NAMESPACE}
-	# Deploy the operator to a separate namespace
-	sed -i "s|{{OPERATOR_NAMESPACE}}|${OPERATOR_NAMESPACE}|g" deploy/kubefledged-operator/deploy/service_account.yaml
-	sed -i "s|{{OPERATOR_NAMESPACE}}|${OPERATOR_NAMESPACE}|g" deploy/kubefledged-operator/deploy/clusterrole_binding.yaml
-	sed -i "s|{{OPERATOR_NAMESPACE}}|${OPERATOR_NAMESPACE}|g" deploy/kubefledged-operator/deploy/operator.yaml
+	# Deploy the operator
+	sed -i "s|{{KUBEFLEDGED_NAMESPACE}}|${KUBEFLEDGED_NAMESPACE}|g" deploy/kubefledged-operator/deploy/service_account.yaml
+	sed -i "s|{{KUBEFLEDGED_NAMESPACE}}|${KUBEFLEDGED_NAMESPACE}|g" deploy/kubefledged-operator/deploy/clusterrole_binding.yaml
+	sed -i "s|{{KUBEFLEDGED_NAMESPACE}}|${KUBEFLEDGED_NAMESPACE}|g" deploy/kubefledged-operator/deploy/operator.yaml
 	kubectl apply -f deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_kubefledgeds_crd.yaml
 	kubectl apply -f deploy/kubefledged-operator/deploy/service_account.yaml
 	kubectl apply -f deploy/kubefledged-operator/deploy/clusterrole.yaml
 	kubectl apply -f deploy/kubefledged-operator/deploy/clusterrole_binding.yaml
 	kubectl apply -f deploy/kubefledged-operator/deploy/operator.yaml
-	# Deploy kube-fledged to a separate namespace
-	sed -i "s|{{OPERATOR_NAMESPACE}}|${OPERATOR_NAMESPACE}|g" deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml
+	# Deploy kube-fledged
 	sed -i "s|{{KUBEFLEDGED_NAMESPACE}}|${KUBEFLEDGED_NAMESPACE}|g" deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml
-	kubectl rollout status deployment kubefledged-operator -n kubefledged-operator --watch
+	kubectl rollout status deployment kubefledged-operator -n ${KUBEFLEDGED_NAMESPACE} --watch
 	kubectl apply -f deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml
 
 update:
@@ -239,17 +233,16 @@ remove-kubefledged:
 	-kubectl delete -f deploy/kubefledged-validatingwebhook.yaml
 
 remove-operator-and-kubefledged:
-	# Remove kubefledged and the namespace
+	# Remove kubefledged
 	-kubectl delete -f deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml
 	-kubectl delete validatingwebhookconfigurations -l app.kubernetes.io/name=kube-fledged
-	-kubectl delete namespace ${KUBEFLEDGED_NAMESPACE}
 	# Remove the kubefledged operator and the namespace
 	-kubectl delete -f deploy/kubefledged-operator/deploy/operator.yaml
 	-kubectl delete -f deploy/kubefledged-operator/deploy/clusterrole_binding.yaml
 	-kubectl delete -f deploy/kubefledged-operator/deploy/clusterrole.yaml
 	-kubectl delete -f deploy/kubefledged-operator/deploy/service_account.yaml
 	-kubectl delete -f deploy/kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_kubefledgeds_crd.yaml
-	-kubectl delete namespace ${OPERATOR_NAMESPACE}
+	-kubectl delete namespace ${KUBEFLEDGED_NAMESPACE}
 	-git checkout deploy/kubefledged-operator/deploy/operator.yaml
 	-git checkout deploy/kubefledged-operator/deploy/clusterrole_binding.yaml
 	-git checkout deploy/kubefledged-operator/deploy/service_account.yaml

diff --git a/README.md b/README.md
@@ -23,6 +23,7 @@ _kube-fledged_ provides CRUD APIs to manage the lifecycle of the image cache, an
 - [Quick Install using YAML manifests](#quick-install-using-yaml-manifests)
 - [Quick Install using Helm chart](#quick-install-using-helm-chart)
 - [Quick Install using Helm operator](#quick-install-using-helm-operator)
+- [Helm chart parameters](#helm-chart-parameters)
 - [Build and Deploy](#build-and-deploy)
   - [Build](#build)
   - [Deploy](#deploy)
@@ -113,7 +114,7 @@ These instructions install _kube-fledged_ to a separate namespace called "kube-f
   $ cd $HOME/src/github.com/senthilrch/kube-fledged
   ```
 
-- Deploy the helm operator to a separate namespace called "kubefledged-operator" and _kube-fledged_ to a separate namespace called "kube-fledged". If you need to deploy to a different namespace, export the variables OPERATOR_NAMESPACE and KUBEFLEDGED_NAMESPACE
+- Deploy the helm operator and _kube-fledged_ to namespace "kube-fledged". If you need to deploy to a different namespace, export the variable KUBEFLEDGED_NAMESPACE
 
   ```
   $ make deploy-using-operator
@@ -126,6 +127,10 @@ These instructions install _kube-fledged_ to a separate namespace called "kube-f
   $ kubectl get imagecaches -n kube-fledged (Output should be: 'No resources found')
   ```
 
+## Helm chart parameters
+
+Parameters of the helm chart are documented [here](docs/helm-parameters.md)
+
 ## Build and Deploy
 
 These instructions will help you build _kube-fledged_ from source and deploy it to a separate namespace called "kube-fledged". If you need to deploy it to a different namespace, edit the namespace field of the manifests in "kube-fledged/deploy" accordingly.

diff --git a/cmd/controller/app/controller.go b/cmd/controller/app/controller.go
@@ -221,30 +221,30 @@ func (c *Controller) Run(threadiness int, stopCh <-chan struct{}) error {
 	defer c.imageworkqueue.ShutDown()
 
 	// Start the informer factories to begin populating the informer caches
-	glog.Info("Starting fledged controller")
+	glog.Info("Starting kubefledged-controller")
 
 	// Wait for the caches to be synced before starting workers
-	glog.Info("Waiting for informer caches to sync")
 	if ok := cache.WaitForCacheSync(stopCh, c.nodesSynced, c.imageCachesSynced); !ok {
 		return fmt.Errorf("failed to wait for caches to sync")
 	}
+	glog.Info("Informer caches synched successfull")
 
-	glog.Info("Starting image cache worker")
 	// Launch workers to process ImageCache resources
 	for i := 0; i < threadiness; i++ {
 		go wait.Until(c.runWorker, time.Second, stopCh)
 	}
+	glog.Info("Image cache worker started")
 
 	if c.imageCacheRefreshFrequency.Nanoseconds() != int64(0) {
-		glog.Info("Starting cache refresh worker")
 		go wait.Until(c.runRefreshWorker, c.imageCacheRefreshFrequency, stopCh)
+		glog.Info("Image cache refresh worker started")
 	}
 
-	glog.Info("Started workers")
 	c.imageManager.Run(stopCh)
 	if err := c.imageManager.Run(stopCh); err != nil {
 		glog.Fatalf("Error running image manager: %s", err.Error())
 	}
+	glog.Info("Image manager started")
 
 	<-stopCh
 	glog.Info("Shutting down workers")
@@ -577,7 +577,7 @@ func (c *Controller) syncHandler(wqKey images.WorkQueueKey) error {
 					status.Message = v1alpha2.ImageCacheMessageImagesPulledSuccessfully
 				}
 			}
-			if v.Status == images.ImageWorkResultStatusFailed && !failures {
+			if (v.Status == images.ImageWorkResultStatusFailed || v.Status == images.ImageWorkResultStatusUnknown) && !failures {
 				failures = true
 				status.Status = v1alpha2.ImageCacheActionStatusFailed
 				if v.ImageWorkRequest.WorkType == images.ImageCachePurge {
@@ -586,7 +586,7 @@ func (c *Controller) syncHandler(wqKey images.WorkQueueKey) error {
 					status.Message = v1alpha2.ImageCacheMessageImagePullFailedForSomeImages
 				}
 			}
-			if v.Status == images.ImageWorkResultStatusFailed {
+			if v.Status == images.ImageWorkResultStatusFailed || v.Status == images.ImageWorkResultStatusUnknown {
 				status.Failures[v.ImageWorkRequest.Image] = append(
 					status.Failures[v.ImageWorkRequest.Image], v1alpha2.NodeReasonMessage{
 						Node:    v.ImageWorkRequest.Node.Labels["kubernetes.io/hostname"],

diff --git a/deploy/kubefledged-deployment-controller.yaml b/deploy/kubefledged-deployment-controller.yaml
@@ -19,7 +19,7 @@ spec:
     spec:
       initContainers:
         - name: wait
-          image: senthilrch/kubefledged-cri-client:v0.8.1
+          image: senthilrch/kubefledged-cri-client:v0.8.2
           command: ["curl"]
           args:
             - "--insecure"
@@ -31,7 +31,7 @@ spec:
             - "https://kubefledged-webhook-server.kube-fledged.svc:3443/readyz"
           imagePullPolicy: Always
       containers:
-      - image: senthilrch/kubefledged-controller:v0.8.1
+      - image: senthilrch/kubefledged-controller:v0.8.2
         command: ["/opt/bin/kubefledged-controller"]
         args:
         - "--stderrthreshold=INFO"
@@ -46,7 +46,7 @@ spec:
             fieldRef:
               fieldPath: metadata.namespace
         - name: KUBEFLEDGED_CRI_CLIENT_IMAGE
-          value: "senthilrch/kubefledged-cri-client:v0.8.1"
+          value: "senthilrch/kubefledged-cri-client:v0.8.2"
         - name: BUSYBOX_IMAGE
           value: "gcr.io/google-containers/busybox:1.27.2"
       serviceAccountName: kubefledged-controller
diff --git a/deploy/kubefledged-deployment-webhook-server.yaml b/deploy/kubefledged-deployment-webhook-server.yaml
@@ -18,7 +18,7 @@ spec:
         app: kubefledged
     spec:
       initContainers:
-      - image: senthilrch/kubefledged-webhook-server:v0.8.1
+      - image: senthilrch/kubefledged-webhook-server:v0.8.2
         command: ["/opt/bin/kubefledged-webhook-server"]
         args:
         - "--stderrthreshold=INFO"
@@ -40,7 +40,7 @@ spec:
         - name: certkey-volume
           mountPath: "/var/run/secrets/webhook-server"
       containers:
-      - image: senthilrch/kubefledged-webhook-server:v0.8.1
+      - image: senthilrch/kubefledged-webhook-server:v0.8.2
         command: ["/opt/bin/kubefledged-webhook-server"]
         args:
         - "--stderrthreshold=INFO"

diff --git a/deploy/kubefledged-operator/deploy/clusterrole_binding.yaml b/deploy/kubefledged-operator/deploy/clusterrole_binding.yaml
@@ -5,7 +5,7 @@ metadata:
 subjects:
 - kind: ServiceAccount
   name: kubefledged-operator
-  namespace: {{OPERATOR_NAMESPACE}}
+  namespace: {{KUBEFLEDGED_NAMESPACE}}
 roleRef:
   kind: ClusterRole
   name: kubefledged-operator

diff --git a/.../kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml b/.../kubefledged-operator/deploy/crds/charts.helm.kubefledged.io_v1alpha2_kubefledged_cr.yaml
@@ -2,7 +2,24 @@ apiVersion: charts.helm.kubefledged.io/v1alpha2
 kind: KubeFledged
 metadata:
   name: kube-fledged
-  namespace: kubefledged-operator
+  namespace: {{KUBEFLEDGED_NAMESPACE}}
 spec:
   # Defaults defined in <project_dir>/helm-charts/kubefledged/values.yaml
-  kubefledgedNameSpace: kube-fledged
+  controllerReplicaCount: 1
+  webhookServerReplicaCount: 1
+  image:
+    kubefledgedControllerRepository: docker.io/senthilrch/kubefledged-controller
+    kubefledgedCRIClientRepository: docker.io/senthilrch/kubefledged-cri-client
+    kubefledgedWebhookServerRepository: docker.io/senthilrch/kubefledged-webhook-server
+    pullPolicy: Always
+  args:
+    controllerLogLevel: INFO
+    controllerImagePullDeadlineDuration: 5m
+    controllerImageCacheRefreshFrequency: 15m
+    controllerImagePullPolicy: IfNotPresent
+    webhookServerLogLevel: INFO
+    webhookServerCertFile: /var/run/secrets/webhook-server/tls.crt
+    webhookServerKeyFile: /var/run/secrets/webhook-server/tls.key
+    webhookServerPort: 443
+  nameOverride: ""
+  fullnameOverride: ""
diff --git a/deploy/kubefledged-operator/deploy/operator.yaml b/deploy/kubefledged-operator/deploy/operator.yaml
@@ -2,7 +2,7 @@ apiVersion: apps/v1
 kind: Deployment
 metadata:
   name: kubefledged-operator
-  namespace: {{OPERATOR_NAMESPACE}}
+  namespace: {{KUBEFLEDGED_NAMESPACE}}
 spec:
   replicas: 1
   selector:
@@ -17,7 +17,7 @@ spec:
       containers:
         - name: kubefledged-operator
           # Replace this with the built image name
-          image: docker.io/senthilrch/kubefledged-operator:v0.8.1
+          image: docker.io/senthilrch/kubefledged-operator:v0.8.2
           imagePullPolicy: Always
           env:
             - name: WATCH_NAMESPACE

diff --git a/deploy/kubefledged-operator/deploy/service_account.yaml b/deploy/kubefledged-operator/deploy/service_account.yaml
@@ -2,4 +2,4 @@ apiVersion: v1
 kind: ServiceAccount
 metadata:
   name: kubefledged-operator
-  namespace: {{OPERATOR_NAMESPACE}}
+  namespace: {{KUBEFLEDGED_NAMESPACE}}
diff --git a/deploy/kubefledged-operator/helm-charts/kubefledged/Chart.yaml b/deploy/kubefledged-operator/helm-charts/kubefledged/Chart.yaml
@@ -14,8 +14,8 @@ 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.
-version: v0.8.1
+version: v0.8.2
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application.
-appVersion: v0.8.1
+appVersion: v0.8.2
diff --git a/docs/helm-parameters.md b/docs/helm-parameters.md
@@ -0,0 +1,21 @@
+# Helm chart parameters
+
+| Parameter | Default value | Description |
+| --------- | ------------- | ----------- |
+| controllerReplicaCount    | 1        | No. of replicas of kubefledged-controller |
+| webhookServerReplicaCount | 1        | No. of replicas of kubefledged-webhook-server |
+| image.kubefledgedControllerRepository | docker.io/senthilrch/kubefledged-controller | Repository name of kubefledged-controller image |
+| image.kubefledgedCRIClientRepository | docker.io/senthilrch/kubefledged-cri-client | Repository name of kubefledged-cri-client image |
+| image.kubefledgedWebhookServerRepository | docker.io/senthilrch/kubefledged-webhook-server | Repository name of kubefledged-webhook-server image |
+| image.pullPolicy | Always | Image pull policy for kubefledged-controller and kubefledged-webhook-server pods |
+| args.controllerLogLevel | INFO | Log level of kubefledged-controller |
+| args.controllerImagePullDeadlineDuration | 5m | Maximum duration allowed for pulling an image. After this duration, image pull is considered to have failed |
+| args.controllerImageCacheRefreshFrequency | 15m | The image cache is refreshed periodically to ensure the cache is up to date. Setting this flag to "0s" will disable refresh |
+| args.controllerImagePullPolicy | IfNotPresent | Image pull policy for pulling images into and refreshing the cache. Possible values are 'IfNotPresent' and 'Always'. Default value is 'IfNotPresent'. Image with no or ":latest" tag are always pulled |
+| args.webhookServerLogLevel | INFO | Log level of kubefledged-webhook-server |
+| args.webhookServerCertFile | /var/run/secrets/webhook-server/tls.crt | Path of server certificate of kubefledged-webhook-server |
+| args.webhookServerKeyFile | /var/run/secrets/webhook-server/tls.key | Path of server key of kubefledged-webhook-server |
+| args.webhookServerPort | 443 | Listening port of kubefledged-webhook-server |
+| nameOverride | "" | nameOverride replaces the name of the chart in Chart.yaml, when this is used to construct Kubernetes object names |
+| fullnameOverride | "" | fullnameOverride completely replaces the generated name |
+|  |  |  |
diff --git a/pkg/images/image_manager.go b/pkg/images/image_manager.go
@@ -53,6 +53,8 @@ const (
 	ImageWorkResultStatusJobCreated = "jobcreated"
 	//ImageWorkResultStatusAlreadyPulled  means image is already present in the node
 	ImageWorkResultStatusAlreadyPulled = "alreadypulled"
+	//ImageWorkResultStatusUnknown  means status of image pull/delete unknown
+	ImageWorkResultStatusUnknown = "unknown"
 )
 
 // ImageManager provides the functionalities for pulling and deleting images
@@ -208,52 +210,61 @@ func (m *ImageManager) updatePendingImageWorkResults(imageCacheName string) erro
 					glog.Errorf("Error listing Pods: %v", err)
 					return err
 				}
-				if len(pods) == 0 {
-					glog.Errorf("No pods matched job %s", job)
-					return fmt.Errorf("no pods matched job %s", job)
-				}
 				if len(pods) > 1 {
 					glog.Errorf("More than one pod matched job %s", job)
 					return fmt.Errorf("more than one pod matched job %s", job)
 				}
-				iwres.Status = ImageWorkResultStatusFailed
-				if iwres.ImageWorkRequest.WorkType == ImageCachePurge {
-					glog.Infof("Job %s expired (delete: %s --> %s)", job, iwres.ImageWorkRequest.Image, iwres.ImageWorkRequest.Node.Labels["kubernetes.io/hostname"])
-				} else {
-					glog.Infof("Job %s expired (pull: %s --> %s)", job, iwres.ImageWorkRequest.Image, iwres.ImageWorkRequest.Node.Labels["kubernetes.io/hostname"])
-				}
-				if pods[0].Status.Phase == corev1.PodPending {
-					if len(pods[0].Status.ContainerStatuses) == 1 {
-						if pods[0].Status.ContainerStatuses[0].State.Waiting != nil {
-							iwres.Reason = pods[0].Status.ContainerStatuses[0].State.Waiting.Reason
-							iwres.Message = pods[0].Status.ContainerStatuses[0].State.Waiting.Message
-						}
-						if pods[0].Status.ContainerStatuses[0].State.Terminated != nil {
-							iwres.Reason = pods[0].Status.ContainerStatuses[0].State.Terminated.Reason
-							iwres.Message = pods[0].Status.ContainerStatuses[0].State.Terminated.Message
-						}
+				if len(pods) == 0 {
+					glog.Warningf("No pods matched job %s", job)
+					if iwres.ImageWorkRequest.WorkType == ImageCachePurge {
+						glog.Warningf("Job %s status unknown (delete: %s --> %s)", job, iwres.ImageWorkRequest.Image, iwres.ImageWorkRequest.Node.Labels["kubernetes.io/hostname"])
 					} else {
-						iwres.Reason = "Pending"
-						iwres.Message = "Check if node is ready"
+						glog.Warningf("Job %s status unknown (pull: %s --> %s)", job, iwres.ImageWorkRequest.Image, iwres.ImageWorkRequest.Node.Labels["kubernetes.io/hostname"])
 					}
+					iwres.Status = ImageWorkResultStatusUnknown
+					iwres.Reason = fmt.Sprintf("No pods matched job %s", job)
+					iwres.Message = fmt.Sprintf("No pods matched job %s", job)
 				}
-				if iwres.ImageWorkRequest.WorkType != ImageCachePurge {
-					fieldSelector := fields.Set{
-						"involvedObject.kind":      "Pod",
-						"involvedObject.name":      pods[0].Name,
-						"involvedObject.namespace": m.fledgedNameSpace,
-						"reason":                   "Failed",
-					}.AsSelector().String()
-
-					eventlist, err := m.kubeclientset.CoreV1().Events(m.fledgedNameSpace).
-						List(context.TODO(), metav1.ListOptions{FieldSelector: fieldSelector})
-					if err != nil {
-						glog.Errorf("Error listing events for pod (%s): %v", pods[0].Name, err)
-						return err
+				if len(pods) == 1 {
+					iwres.Status = ImageWorkResultStatusFailed
+					if iwres.ImageWorkRequest.WorkType == ImageCachePurge {
+						glog.Infof("Job %s expired (delete: %s --> %s)", job, iwres.ImageWorkRequest.Image, iwres.ImageWorkRequest.Node.Labels["kubernetes.io/hostname"])
+					} else {
+						glog.Infof("Job %s expired (pull: %s --> %s)", job, iwres.ImageWorkRequest.Image, iwres.ImageWorkRequest.Node.Labels["kubernetes.io/hostname"])
+					}
+					if pods[0].Status.Phase == corev1.PodPending {
+						if len(pods[0].Status.ContainerStatuses) == 1 {
+							if pods[0].Status.ContainerStatuses[0].State.Waiting != nil {
+								iwres.Reason = pods[0].Status.ContainerStatuses[0].State.Waiting.Reason
+								iwres.Message = pods[0].Status.ContainerStatuses[0].State.Waiting.Message
+							}
+							if pods[0].Status.ContainerStatuses[0].State.Terminated != nil {
+								iwres.Reason = pods[0].Status.ContainerStatuses[0].State.Terminated.Reason
+								iwres.Message = pods[0].Status.ContainerStatuses[0].State.Terminated.Message
+							}
+						} else {
+							iwres.Reason = "Pending"
+							iwres.Message = "Check if node is ready"
+						}
 					}
+					if iwres.ImageWorkRequest.WorkType != ImageCachePurge {
+						fieldSelector := fields.Set{
+							"involvedObject.kind":      "Pod",
+							"involvedObject.name":      pods[0].Name,
+							"involvedObject.namespace": m.fledgedNameSpace,
+							"reason":                   "Failed",
+						}.AsSelector().String()
+
+						eventlist, err := m.kubeclientset.CoreV1().Events(m.fledgedNameSpace).
+							List(context.TODO(), metav1.ListOptions{FieldSelector: fieldSelector})
+						if err != nil {
+							glog.Errorf("Error listing events for pod (%s): %v", pods[0].Name, err)
+							return err
+						}
 
-					for _, v := range eventlist.Items {
-						iwres.Message = iwres.Message + ":" + v.Message
+						for _, v := range eventlist.Items {
+							iwres.Message = iwres.Message + ":" + v.Message
+						}
 					}
 				}
 				m.imageworkstatus[job] = iwres