nfd-worker: drop the gRPC health port

To be replaced with plain http.

Author: marquiz

cmd/nfd-worker/main.go

@@ -110,8 +110,6 @@ func initFlags(flagset *flag.FlagSet) (*worker.Args, *worker.ConfigOverrideArgs)
 		"Do not publish feature labels")
 	flagset.IntVar(&args.MetricsPort, "metrics", 8081,
 		"Port on which to expose metrics.")
-	flagset.IntVar(&args.GrpcHealthPort, "grpc-health", 8082,
-		"Port on which to expose the grpc health endpoint.")
 	flagset.StringVar(&args.Options, "options", "",
 		"Specify config options from command line. Config options are specified "+
 			"in the same format as in the config file (i.e. json or yaml). These options")

deployment/base/worker-daemonset/worker-daemonset.yaml

@@ -20,13 +20,9 @@ spec:
           image: gcr.io/k8s-staging-nfd/node-feature-discovery:master
           imagePullPolicy: Always
           livenessProbe:
-            grpc:
-              port: 8082
             initialDelaySeconds: 10
             periodSeconds: 10
           readinessProbe:
-            grpc:
-              port: 8082
             initialDelaySeconds: 5
             periodSeconds: 10
             failureThreshold: 10

deployment/helm/node-feature-discovery/templates/worker.yaml

@@ -47,8 +47,6 @@ spec:
         image: "{{ .Values.image.repository }}:{{ .Values.image.tag | default .Chart.AppVersion }}"
         imagePullPolicy: {{ .Values.image.pullPolicy }}
         livenessProbe:
-          grpc:
-            port: {{ .Values.worker.healthPort | default "8082" }}
         {{- with .Values.worker.livenessProbe.initialDelaySeconds }}
           initialDelaySeconds: {{ . }}
         {{- end }}
@@ -62,8 +60,6 @@ spec:
           timeoutSeconds: {{ . }}
         {{- end }}
         readinessProbe:
-          grpc:
-            port: {{ .Values.worker.healthPort | default "8082" }}
         {{- with .Values.worker.readinessProbe.initialDelaySeconds }}
           initialDelaySeconds: {{ . }}
         {{- end }}
@@ -105,15 +101,12 @@ spec:
         - "-feature-gates={{ $key }}={{ $value }}"
         {{- end }}
         - "-metrics={{ .Values.worker.metricsPort | default "8081"}}"
-        - "-grpc-health={{ .Values.worker.healthPort | default "8082" }}"
         {{- with .Values.worker.extraArgs }}
         {{- toYaml . | nindent 8 }}
         {{- end }}
         ports:
           - containerPort: {{ .Values.worker.metricsPort | default "8081"}}
             name: metrics
-          - containerPort: {{ .Values.worker.healthPort | default "8082" }}
-            name: health
         volumeMounts:
         - name: host-boot
           mountPath: "/host-boot"

deployment/helm/node-feature-discovery/values.yaml

@@ -419,7 +419,6 @@ worker:
 ### <NFD-WORKER-CONF-END-DO-NOT-REMOVE>
 
   metricsPort: 8081
-  healthPort: 8082
   daemonsetAnnotations: {}
   podSecurityContext: {}
     # fsGroup: 2000
@@ -433,15 +432,11 @@ worker:
     # runAsUser: 1000
 
   livenessProbe:
-    grpc:
-      port: 8082
     initialDelaySeconds: 10
     # failureThreshold: 3
     # periodSeconds: 10
     # timeoutSeconds: 1
   readinessProbe:
-    grpc:
-      port: 8082
     initialDelaySeconds: 5
     failureThreshold: 10
     # periodSeconds: 10

docs/deployment/helm.md

@@ -234,7 +234,6 @@ API's you need to install the prometheus operator in your cluster.
 | `worker.enable`                             | bool    | true                    | Specifies whether nfd-worker should be deployed                                                                                                                                                              |
 | `worker.hostNetwork`                        | bool    | false                   | Specifies whether to enable or disable running the container in the host's network namespace                                                                                                                 |
 | `worker.metricsPort`                        | int     | 8081                    | Port on which to expose metrics from components to prometheus operator. **DEPRECATED**: will be replaced by `worker.port` in NFD v0.18.                                                                      |
-| `worker.healthPort`                         | int     | 8082                    | Port on which to expose the grpc health endpoint, will be also used for the probes. **DEPRECATED**: will be replaced by `worker.port` in NFD v0.18.                                                          |
 | `worker.config`                             | dict    |                         | NFD worker [configuration](../reference/worker-configuration-reference)                                                                                                                                      |
 | `worker.podSecurityContext`                 | dict    | {}                      | [PodSecurityContext](https://kubernetes.io/docs/tasks/configure-pod-container/security-context/#set-the-security-context-for-a-pod) holds pod-level security attributes and common container settins         |
 | `worker.securityContext`                    | dict    | {}                      | Container [security settings](https://kubernetes.io/docs/tasks/configure-pod-container/security-context/#set-the-security-context-for-a-container)                                                           |

pkg/nfd-worker/nfd-worker.go

@@ -19,7 +19,6 @@ package nfdworker
 import (
 	"encoding/json"
 	"fmt"
-	"net"
 	"os"
 	"path/filepath"
 	"regexp"
@@ -29,9 +28,6 @@ import (
 
 	"golang.org/x/exp/maps"
 	"golang.org/x/net/context"
-	"google.golang.org/grpc"
-	"google.golang.org/grpc/health"
-	"google.golang.org/grpc/health/grpc_health_v1"
 	"k8s.io/apimachinery/pkg/api/errors"
 	"k8s.io/apimachinery/pkg/types"
 	"k8s.io/apimachinery/pkg/util/validation"
@@ -93,14 +89,13 @@ type Labels map[string]string
 
 // Args are the command line arguments of NfdWorker.
 type Args struct {
-	ConfigFile     string
-	Klog           map[string]*utils.KlogFlagVal
-	Kubeconfig     string
-	Oneshot        bool
-	Options        string
-	MetricsPort    int
-	GrpcHealthPort int
-	NoOwnerRefs    bool
+	ConfigFile  string
+	Klog        map[string]*utils.KlogFlagVal
+	Kubeconfig  string
+	Oneshot     bool
+	Options     string
+	MetricsPort int
+	NoOwnerRefs bool
 
 	Overrides ConfigOverrideArgs
 }
@@ -118,7 +113,6 @@ type nfdWorker struct {
 	configFilePath      string
 	config              *NFDConfig
 	kubernetesNamespace string
-	healthServer        *grpc.Server
 	k8sClient           k8sclient.Interface
 	nfdClient           nfdclient.Interface
 	stop                chan struct{} // channel for signaling stop
@@ -217,29 +211,6 @@ func (i *infiniteTicker) Reset(d time.Duration) {
 	}
 }
 
-func (w *nfdWorker) startGrpcHealthServer(errChan chan<- error) error {
-	lis, err := net.Listen("tcp", fmt.Sprintf(":%d", w.args.GrpcHealthPort))
-	if err != nil {
-		return fmt.Errorf("failed to listen: %w", err)
-	}
-
-	s := grpc.NewServer()
-	grpc_health_v1.RegisterHealthServer(s, health.NewServer())
-	klog.InfoS("gRPC health server serving", "port", w.args.GrpcHealthPort)
-
-	go func() {
-		defer func() {
-			lis.Close()
-		}()
-		if err := s.Serve(lis); err != nil {
-			errChan <- fmt.Errorf("gRPC health server exited with an error: %w", err)
-		}
-		klog.InfoS("gRPC health server stopped")
-	}()
-	w.healthServer = s
-	return nil
-}
-
 // Run feature discovery.
 func (w *nfdWorker) runFeatureDiscovery() error {
 	discoveryStart := time.Now()
@@ -344,20 +315,10 @@ func (w *nfdWorker) Run() error {
 		return nil
 	}
 
-	grpcErr := make(chan error)
-
-	// Start gRPC server for liveness probe (at this point we're "live")
-	if w.args.GrpcHealthPort != 0 {
-		if err := w.startGrpcHealthServer(grpcErr); err != nil {
-			return fmt.Errorf("failed to start gRPC health server: %w", err)
-		}
-	}
+	// Start readiness probe (at this point we're "ready and live")
 
 	for {
 		select {
-		case err := <-grpcErr:
-			return fmt.Errorf("error in serving gRPC: %w", err)
-
 		case <-labelTrigger.C:
 			err = w.runFeatureDiscovery()
 			if err != nil {
@@ -366,9 +327,6 @@ func (w *nfdWorker) Run() error {
 
 		case <-w.stop:
 			klog.InfoS("shutting down nfd-worker")
-			if w.healthServer != nil {
-				w.healthServer.GracefulStop()
-			}
 			return nil
 		}
 	}