scheduler server add leader elect

Signed-off-by: rongfu.leng <[email protected]>
Project-HAMi · Apr 12, 2024 · a55741e · a55741e
1 parent 8381474
commit a55741e
Show file tree

Hide file tree

Showing 3 changed files with 10 additions and 3 deletions.
diff --git a/charts/hami/templates/scheduler/deployment.yaml b/charts/hami/templates/scheduler/deployment.yaml
@@ -47,6 +47,9 @@ spec:
             - {{ . }}
             {{- end }}
             {{- end }}
+            - --leader-elect={{ .Values.scheduler.leaderElect }}
+            - --leader-elect-resource-name={{ .Values.schedulerName }}
+            - --leader-elect-resource-namespace={{ .Release.Namespace }}
           volumeMounts:
             - name: scheduler-config
               mountPath: /config
@@ -71,6 +74,9 @@ spec:
             - --default-cores={{ .Values.scheduler.defaultCores }}
             - --iluvatar-memory={{ .Values.iluvatarResourceMem }}
             - --iluvatar-cores={{ .Values.iluvatarResourceCore }}
+            - --leader-elect={{ .Values.scheduler.leaderElect }}
+            - --leader-elect-resource-name={{ .Values.schedulerName }}
+            - --leader-elect-resource-namespace={{ .Release.Namespace }}
             {{- range .Values.scheduler.extender.extraArgs }}
             - {{ . }}
             {{- end }}

diff --git a/charts/hami/values.yaml b/charts/hami/values.yaml
@@ -2,7 +2,7 @@
 
 nameOverride: ""
 fullnameOverride: ""
-imagePullSecrets: []
+imagePullSecrets: [ ]
 version: "v2.3.9"
 
 #Nvidia GPU Parameters
@@ -45,6 +45,7 @@ scheduler:
   defaultCores: 0
   defaultGPUNum: 1
   metricsBindAddress: ":9395"
+  leaderElect: true
   kubeScheduler:
     # @param enabled indicate whether to run kube-scheduler container in the scheduler pod, it's true by default.
     enabled: true
@@ -56,7 +57,6 @@ scheduler:
       - -v=4
     extraArgs:
       - --policy-config-file=/config/config.json
-      - --leader-elect=false
       - -v=4
   extender:
     image: "projecthami/hami"

diff --git a/cmd/scheduler/main.go b/cmd/scheduler/main.go
@@ -57,6 +57,7 @@ func init() {
 	rootCmd.Flags().Int32Var(&config.DefaultCores, "default-cores", 0, "default gpu core percentage to allocate")
 	rootCmd.Flags().Int32Var(&config.DefaultResourceNum, "default-gpu", 1, "default gpu to allocate")
 	rootCmd.Flags().StringVar(&config.MetricsBindAddress, "metrics-bind-address", ":9395", "The TCP address that the scheduler should bind to for serving prometheus metrics(e.g. 127.0.0.1:9395, :9395)")
+
 	rootCmd.PersistentFlags().AddGoFlagSet(device.GlobalFlagSet())
 	rootCmd.AddCommand(version.VersionCmd)
 	rootCmd.Flags().AddGoFlagSet(util.InitKlogFlags())
@@ -65,7 +66,7 @@ func init() {
 func start() {
 	sher = scheduler.NewScheduler()
 	sher.Start()
-	defer sher.Stop()
+	go sher.Stop()
 
 	// start monitor metrics
 	go sher.RegisterFromNodeAnnotations()