Merge pull request #246 from arangodb/bugfix/resilience-improvements

Resilience improvements

Author: ewoutp

examples/metrics/dashboard.json

@@ -1,34 +1,22 @@
 # This example shows how to integrate with the Prometheus Operator
 # to bring metrics from kube-arangodb to Prometheus.
 
-apiVersion: v1
-kind: Service
-metadata:
-  name: arango-deployment-operator
-  labels:
-    app: arango-deployment-operator
-spec:
-  selector:
-    app: arango-deployment-operator
-  ports:
-  - name: metrics
-    port: 8528
-
----
-
 apiVersion: monitoring.coreos.com/v1
 kind: ServiceMonitor
 metadata:
   name: arango-deployment-operator
+  namespace: monitoring
   labels:
-    team: frontend
+    prometheus: kube-prometheus
 spec:
   selector:
     matchLabels:
       app: arango-deployment-operator
+  namespaceSelector:
+    matchNames:
+    - default
   endpoints:
-  - port: metrics
+  - port: server
     scheme: https
     tlsConfig:
       insecureSkipVerify: true
-

examples/metrics/deployment-operator-servicemonitor.yaml

@@ -147,4 +147,6 @@ func cmdLifecycleCopyRun(cmd *cobra.Command, args []string) {
 	if err := os.Chmod(targetPath, 0755); err != nil {
 		cliLog.Fatal().Err(err).Msg("Failed to chmod")
 	}
+
+	cliLog.Info().Msgf("Executable copied to %s", targetPath)
 }

lifecycle.go

@@ -40,6 +40,10 @@ const (
 	// ConditionTypeCleanedOut indicates that the member (dbserver) has been cleaned out.
 	// Always check in combination with ConditionTypeTerminated.
 	ConditionTypeCleanedOut ConditionType = "CleanedOut"
+	// ConditionTypeAgentRecoveryNeeded indicates that the member (agent) will no
+	// longer recover from its current volume and there has to be rebuild
+	// using the recovery procedure.
+	ConditionTypeAgentRecoveryNeeded ConditionType = "AgentRecoveryNeeded"
 	// ConditionTypePodSchedulingFailure indicates that one or more pods belonging to the deployment cannot be schedule.
 	ConditionTypePodSchedulingFailure ConditionType = "PodSchedulingFailure"
 	// ConditionTypeSecretsChanged indicates that the value of one of more secrets used by

pkg/apis/deployment/v1alpha/conditions.go

@@ -24,6 +24,7 @@ package v1alpha
 
 import (
 	"math/rand"
+	"sort"
 
 	"github.com/pkg/errors"
 )
@@ -83,7 +84,9 @@ func (l *MemberStatusList) add(m MemberStatus) error {
 			return maskAny(errors.Wrapf(AlreadyExistsError, "Member '%s' already exists", m.ID))
 		}
 	}
-	*l = append(src, m)
+	newList := append(src, m)
+	sort.Slice(newList, func(i, j int) bool { return newList[i].ID < newList[j].ID })
+	*l = newList
 	return nil
 }
 

pkg/apis/deployment/v1alpha/member_status_list.go

@@ -108,15 +108,15 @@ func (d *Deployment) ensureAccessPackage(apSecretName string) error {
 
 	// Fetch client authentication CA
 	clientAuthSecretName := spec.Sync.Authentication.GetClientCASecretName()
-	clientAuthCert, clientAuthKey, _, err := k8sutil.GetCASecret(d.deps.KubeCli.CoreV1(), clientAuthSecretName, ns, nil)
+	clientAuthCert, clientAuthKey, _, err := k8sutil.GetCASecret(secrets, clientAuthSecretName, nil)
 	if err != nil {
 		log.Debug().Err(err).Msg("Failed to get client-auth CA secret")
 		return maskAny(err)
 	}
 
 	// Fetch TLS CA public key
 	tlsCASecretName := spec.Sync.TLS.GetCASecretName()
-	tlsCACert, err := k8sutil.GetCACertficateSecret(d.deps.KubeCli.CoreV1(), tlsCASecretName, ns)
+	tlsCACert, err := k8sutil.GetCACertficateSecret(secrets, tlsCASecretName)
 	if err != nil {
 		log.Debug().Err(err).Msg("Failed to get TLS CA secret")
 		return maskAny(err)

pkg/deployment/access_package.go

@@ -171,8 +171,9 @@ func (d *Deployment) GetSyncServerClient(ctx context.Context, group api.ServerGr
 	log := d.deps.Log
 	kubecli := d.deps.KubeCli
 	ns := d.apiObject.GetNamespace()
+	secrets := kubecli.CoreV1().Secrets(ns)
 	secretName := d.apiObject.Spec.Sync.Monitoring.GetTokenSecretName()
-	monitoringToken, err := k8sutil.GetTokenSecret(kubecli.CoreV1(), secretName, ns)
+	monitoringToken, err := k8sutil.GetTokenSecret(secrets, secretName)
 	if err != nil {
 		log.Debug().Err(err).Str("secret-name", secretName).Msg("Failed to get sync monitoring secret")
 		return nil, maskAny(err)
@@ -331,7 +332,8 @@ func (d *Deployment) GetPvc(pvcName string) (*v1.PersistentVolumeClaim, error) {
 func (d *Deployment) GetTLSKeyfile(group api.ServerGroup, member api.MemberStatus) (string, error) {
 	secretName := k8sutil.CreateTLSKeyfileSecretName(d.apiObject.GetName(), group.AsRole(), member.ID)
 	ns := d.apiObject.GetNamespace()
-	result, err := k8sutil.GetTLSKeyfileSecret(d.deps.KubeCli.CoreV1(), secretName, ns)
+	secrets := d.deps.KubeCli.CoreV1().Secrets(ns)
+	result, err := k8sutil.GetTLSKeyfileSecret(secrets, secretName)
 	if err != nil {
 		return "", maskAny(err)
 	}
@@ -353,8 +355,9 @@ func (d *Deployment) DeleteTLSKeyfile(group api.ServerGroup, member api.MemberSt
 // Returns: publicKey, privateKey, ownerByDeployment, error
 func (d *Deployment) GetTLSCA(secretName string) (string, string, bool, error) {
 	ns := d.apiObject.GetNamespace()
+	secrets := d.deps.KubeCli.CoreV1().Secrets(ns)
 	owner := d.apiObject.AsOwner()
-	cert, priv, isOwned, err := k8sutil.GetCASecret(d.deps.KubeCli.CoreV1(), secretName, ns, &owner)
+	cert, priv, isOwned, err := k8sutil.GetCASecret(secrets, secretName, &owner)
 	if err != nil {
 		return "", "", false, maskAny(err)
 	}

pkg/deployment/context_impl.go

@@ -42,6 +42,7 @@ import (
 	"github.com/arangodb/kube-arangodb/pkg/deployment/resilience"
 	"github.com/arangodb/kube-arangodb/pkg/deployment/resources"
 	"github.com/arangodb/kube-arangodb/pkg/generated/clientset/versioned"
+	"github.com/arangodb/kube-arangodb/pkg/util"
 	"github.com/arangodb/kube-arangodb/pkg/util/k8sutil"
 	"github.com/arangodb/kube-arangodb/pkg/util/retry"
 	"github.com/arangodb/kube-arangodb/pkg/util/trigger"
@@ -78,8 +79,8 @@ type deploymentEvent struct {
 
 const (
 	deploymentEventQueueSize = 256
-	minInspectionInterval    = time.Second // Ensure we inspect the generated resources no less than with this interval
-	maxInspectionInterval    = time.Minute // Ensure we inspect the generated resources no less than with this interval
+	minInspectionInterval    = util.Interval(time.Second) // Ensure we inspect the generated resources no less than with this interval
+	maxInspectionInterval    = util.Interval(time.Minute) // Ensure we inspect the generated resources no less than with this interval
 )
 
 // Deployment is the in process state of an ArangoDeployment.
@@ -140,6 +141,7 @@ func New(config Config, deps Dependencies, apiObject *api.ArangoDeployment) (*De
 		ci := newClusterScalingIntegration(d)
 		d.clusterScalingIntegration = ci
 		go ci.ListenForClusterEvents(d.stopCh)
+		go d.resources.RunDeploymentHealthLoop(d.stopCh)
 	}
 	if config.AllowChaos {
 		d.chaosMonkey = chaos.NewMonkey(deps.Log, d)
@@ -247,21 +249,21 @@ func (d *Deployment) run() {
 			}
 
 		case <-d.inspectTrigger.Done():
+			log.Debug().Msg("Inspect deployment...")
 			inspectionInterval = d.inspectDeployment(inspectionInterval)
+			log.Debug().Str("interval", inspectionInterval.String()).Msg("...inspected deployment")
 
 		case <-d.updateDeploymentTrigger.Done():
+			inspectionInterval = minInspectionInterval
 			if err := d.handleArangoDeploymentUpdatedEvent(); err != nil {
 				d.CreateEvent(k8sutil.NewErrorEvent("Failed to handle deployment update", err, d.GetAPIObject()))
 			}
 
-		case <-time.After(inspectionInterval):
+		case <-inspectionInterval.After():
 			// Trigger inspection
 			d.inspectTrigger.Trigger()
 			// Backoff with next interval
-			inspectionInterval = time.Duration(float64(inspectionInterval) * 1.5)
-			if inspectionInterval > maxInspectionInterval {
-				inspectionInterval = maxInspectionInterval
-			}
+			inspectionInterval = inspectionInterval.Backoff(1.5, maxInspectionInterval)
 		}
 	}
 }

pkg/deployment/deployment.go

@@ -27,26 +27,35 @@ import (
 	"time"
 
 	api "github.com/arangodb/kube-arangodb/pkg/apis/deployment/v1alpha"
+	"github.com/arangodb/kube-arangodb/pkg/metrics"
+	"github.com/arangodb/kube-arangodb/pkg/util"
 	"github.com/arangodb/kube-arangodb/pkg/util/k8sutil"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 )
 
+var (
+	inspectDeploymentDurationGauges = metrics.MustRegisterGaugeVec(metricsComponent, "inspect_deployment_duration", "Amount of time taken by a single inspection of a deployment (in sec)", metrics.DeploymentName)
+)
+
 // inspectDeployment inspects the entire deployment, creates
 // a plan to update if needed and inspects underlying resources.
 // This function should be called when:
 // - the deployment has changed
 // - any of the underlying resources has changed
 // - once in a while
 // Returns the delay until this function should be called again.
-func (d *Deployment) inspectDeployment(lastInterval time.Duration) time.Duration {
+func (d *Deployment) inspectDeployment(lastInterval util.Interval) util.Interval {
 	log := d.deps.Log
+	start := time.Now()
 
 	nextInterval := lastInterval
 	hasError := false
 	ctx := context.Background()
+	deploymentName := d.apiObject.GetName()
+	defer metrics.SetDuration(inspectDeploymentDurationGauges.WithLabelValues(deploymentName), start)
 
 	// Check deployment still exists
-	updated, err := d.deps.DatabaseCRCli.DatabaseV1alpha().ArangoDeployments(d.apiObject.GetNamespace()).Get(d.apiObject.GetName(), metav1.GetOptions{})
+	updated, err := d.deps.DatabaseCRCli.DatabaseV1alpha().ArangoDeployments(d.apiObject.GetNamespace()).Get(deploymentName, metav1.GetOptions{})
 	if k8sutil.IsNotFound(err) {
 		// Deployment is gone
 		log.Info().Msg("Deployment is gone")
@@ -87,13 +96,17 @@ func (d *Deployment) inspectDeployment(lastInterval time.Duration) time.Duration
 		}
 
 		// Inspection of generated resources needed
-		if err := d.resources.InspectPods(ctx); err != nil {
+		if x, err := d.resources.InspectPods(ctx); err != nil {
 			hasError = true
 			d.CreateEvent(k8sutil.NewErrorEvent("Pod inspection failed", err, d.apiObject))
+		} else {
+			nextInterval = nextInterval.ReduceTo(x)
 		}
-		if err := d.resources.InspectPVCs(ctx); err != nil {
+		if x, err := d.resources.InspectPVCs(ctx); err != nil {
 			hasError = true
 			d.CreateEvent(k8sutil.NewErrorEvent("PVC inspection failed", err, d.apiObject))
+		} else {
+			nextInterval = nextInterval.ReduceTo(x)
 		}
 
 		// Check members for resilience
@@ -149,9 +162,11 @@ func (d *Deployment) inspectDeployment(lastInterval time.Duration) time.Duration
 		}
 
 		// At the end of the inspect, we cleanup terminated pods.
-		if err := d.resources.CleanupTerminatedPods(); err != nil {
+		if x, err := d.resources.CleanupTerminatedPods(); err != nil {
 			hasError = true
 			d.CreateEvent(k8sutil.NewErrorEvent("Pod cleanup failed", err, d.apiObject))
+		} else {
+			nextInterval = nextInterval.ReduceTo(x)
 		}
 	}
 
@@ -164,10 +179,7 @@ func (d *Deployment) inspectDeployment(lastInterval time.Duration) time.Duration
 	} else {
 		d.recentInspectionErrors = 0
 	}
-	if nextInterval > maxInspectionInterval {
-		nextInterval = maxInspectionInterval
-	}
-	return nextInterval
+	return nextInterval.ReduceTo(maxInspectionInterval)
 }
 
 // triggerInspection ensures that an inspection is run soon.

pkg/deployment/deployment_inspector.go

@@ -0,0 +1,28 @@
+//
+// DISCLAIMER
+//
+// Copyright 2018 ArangoDB GmbH, Cologne, Germany
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//
+// Copyright holder is ArangoDB GmbH, Cologne, Germany
+//
+// Author Ewout Prangsma
+//
+
+package deployment
+
+const (
+	// Component name for metrics of this package
+	metricsComponent = "deployment"
+)