Use static CPU Manager for Core Pinning

ansible/inventory.yaml

@@ -31,8 +31,8 @@ k3s_cluster:
     api_endpoint: "{{ hostvars[groups['server'][0]]['ansible_host'] | default(groups['server'][0]) }}"
 
     # Optional vars
-    extra_server_args: "--flannel-backend=none --disable-network-policy --write-kubeconfig-mode 644 --kubelet-arg=max-pods=1000 --kubelet-arg=resolv-conf=/etc/rancher/resolv.conf"
-    extra_agent_args: "--kubelet-arg=max-pods=1000 --kubelet-arg=resolv-conf=/etc/rancher/resolv.conf"
+    extra_server_args: "--flannel-backend=none --disable-network-policy --write-kubeconfig-mode 644 --kubelet-arg=max-pods=1000 --kubelet-arg=resolv-conf=/etc/rancher/resolv.conf --kubelet-arg=config=/etc/rancher/kubelet-config.yaml"
+    extra_agent_args: "--kubelet-arg=max-pods=1000 --kubelet-arg=resolv-conf=/etc/rancher/resolv.conf --kubelet-arg=config=/etc/rancher/kubelet-config.yaml"
     # extra_install_envs: { 'INSTALL_K3S_SKIP_SELINUX_RPM': 'true' }
     # cluster_context: vaclab
     # api_port: 6443

ansible/playbooks/tasks/system_tuning.yaml

@@ -65,6 +65,20 @@
       nameserver 1.1.1.1
       nameserver 185.12.64.2
 
+- name: Create kubelet config for static cpuManager
+  ansible.builtin.copy:
+    dest: /etc/rancher/kubelet-config.yaml
+    mode: '0644'
+    content: |
+      kind: KubeletConfiguration
+      apiVersion: kubelet.config.k8s.io/v1beta1
+      cpuManagerPolicy: static
+      cpuManagerReconcilePeriod: 10s
+      kubeReserved:
+        cpu: "1"
+      systemReserved:
+        cpu: "1"
+
 - name: Configure user limits for file descriptors
   ansible.builtin.copy:
     dest: /etc/security/limits.d/99-vaclab-limits.conf

fleet/dst-dashboard/backend/backend.yaml

@@ -30,7 +30,7 @@ spec:
           effect: "PreferNoSchedule"
       containers:
       - name: dst-dashboard-backend
-        image: mamoutoudiarra/dst-dashboard-backend:v1.1
+        image: mamoutoudiarra/dst-dashboard-backend:v1.3
         imagePullPolicy: IfNotPresent
         ports:
         - containerPort: 8000