Add support for GPU accelerator

inventory/sample/group_vars/k8s-cluster/k8s-cluster.yml

@@ -182,6 +182,20 @@ podsecuritypolicy_enabled: false
 ## See https://github.com/kubernetes-incubator/kubespray/issues/2141
 ## Set this variable to true to get rid of this issue
 volume_cross_zone_attachment: false
-
 # Add Persistent Volumes Storage Class for corresponding cloud provider ( OpenStack is only supported now )
 persistent_volumes_enabled: false
+# Add Persistent Volumes Storage Class for corresponding cloud provider ( OpenStack is only supported now )
+persistent_volumes_enabled: false
+
+## Container Engine Acceleration
+## Enable container accelertion feature, for example use gpu acceleration in containers
+# nvidia_accelerator_enabled: true
+## Nvidia GPU driver install. Install will by done by a (init) pod running as a daemonset.
+## Important: if you use Ubuntu then you should set in all.yml 'docker_storage_options: -s overlay2'
+## Array with nvida_gpu_nodes, leave empty or comment if you dont't want to install drivers.
+## Labels and taints won't be set to nodes if they are not in the array.
+# nvidia_gpu_nodes:
+#   - kube-gpu-001
+# nvidia_driver_version: "384.111"
+## flavor can be tesla or gtx
+# nvidia_gpu_flavor: gtx

roles/kubernetes-apps/container_engine_accelerator/meta/main.yml

@@ -0,0 +1,8 @@
+---
+dependencies:
+  - role: kubernetes-apps/container_engine_accelerator/nvidia_gpu
+    when: nvidia_accelerator_enabled
+    tags:
+      - apps
+      - nvidia_gpu
+      - container_engine_accelerator

roles/kubernetes-apps/container_engine_accelerator/nvidia_gpu/defaults/main.yml

@@ -0,0 +1,10 @@
+---
+nvidia_accelerator_enabled: false
+nvidia_driver_version: "390.87"
+nvidia_gpu_tesla_base_url: https://us.download.nvidia.com/tesla/
+nvidia_gpu_gtx_base_url: http://us.download.nvidia.com/XFree86/Linux-x86_64/
+nvidia_gpu_flavor: tesla
+nvidia_url_end: "{{nvidia_driver_version}}/NVIDIA-Linux-x86_64-{{nvidia_driver_version}}.run"
+nvidia_driver_install_container: false
+nvidia_driver_install_supported: false
+nvidia_gpu_nodes: []

roles/kubernetes-apps/container_engine_accelerator/nvidia_gpu/tasks/main.yml

@@ -0,0 +1,54 @@
+---
+
+- name: Container Engine Acceleration Nvidia GPU| gather os specific variables
+  include_vars: "{{ item }}"
+  with_first_found:
+    - files:
+        - "{{ ansible_distribution|lower }}-{{ ansible_distribution_version|lower|replace('/', '_') }}.yml"
+        - "{{ ansible_distribution|lower }}-{{ ansible_distribution_release }}.yml"
+        - "{{ ansible_distribution|lower }}-{{ ansible_distribution_major_version|lower|replace('/', '_') }}.yml"
+        - "{{ ansible_distribution|lower }}.yml"
+        - "{{ ansible_os_family|lower }}.yml"
+      skip: true
+
+- name: Container Engine Acceleration Nvidia GPU | Set fact of download url Tesla
+  set_fact:
+    nvidia_driver_download_url_default: "{{nvidia_gpu_tesla_base_url}}{{nvidia_url_end}}"
+  when: nvidia_gpu_flavor|lower == "tesla"
+
+- name: Container Engine Acceleration Nvidia GPU | Set fact of download url GTX
+  set_fact:
+    nvidia_driver_download_url_default: "{{nvidia_gpu_gtx_base_url}}{{nvidia_url_end}}"
+  when: nvidia_gpu_flavor|lower == "gtx"
+
+- name: Container Engine Acceleration Nvidia GPU | Create addon dir
+  file:
+    path: "{{ kube_config_dir }}/addons/container_engine_accelerator"
+    owner: root
+    group: root
+    mode: 0755
+    recurse: true
+
+- name: Container Engine Acceleration Nvidia GPU | Create manifests for nvidia accelerators
+  template:
+    src: "{{ item.file }}.j2"
+    dest: "{{ kube_config_dir }}/addons/container_engine_accelerator/{{ item.file }}"
+  with_items:
+    - { name: nvidia-driver-install-daemonset, file: nvidia-driver-install-daemonset.yml, type: daemonset }
+    - { name: k8s-device-plugin-nvidia-daemonset, file: k8s-device-plugin-nvidia-daemonset.yml, type: daemonset }
+  register: container_engine_accelerator_manifests
+  when:
+    - inventory_hostname == groups['kube-master'][0] and nvidia_driver_install_container
+
+- name: Container Engine Acceleration Nvidia GPU | Apply manifests for nvidia accelerators
+  kube:
+    name: "{{ item.item.name }}"
+    namespace: "kube-system"
+    kubectl: "{{ bin_dir }}/kubectl"
+    resource: "{{ item.item.type }}"
+    filename: "{{ kube_config_dir }}/addons/container_engine_accelerator/{{ item.item.file }}"
+    state: "latest"
+  with_items:
+    - "{{container_engine_accelerator_manifests.results}}"
+  when:
+    - inventory_hostname == groups['kube-master'][0] and nvidia_driver_install_container and nvidia_driver_install_supported

roles/kubernetes-apps/container_engine_accelerator/nvidia_gpu/templates/k8s-device-plugin-nvidia-daemonset.yml.j2

@@ -0,0 +1,61 @@
+apiVersion: apps/v1
+kind: DaemonSet
+metadata:
+  name: nvidia-gpu-device-plugin
+  namespace: kube-system
+  labels:
+    k8s-app: nvidia-gpu-device-plugin
+    addonmanager.kubernetes.io/mode: Reconcile
+spec:
+  selector:
+    matchLabels:
+      k8s-app: nvidia-gpu-device-plugin
+  template:
+    metadata:
+      labels:
+        k8s-app: nvidia-gpu-device-plugin
+      annotations:
+        scheduler.alpha.kubernetes.io/critical-pod: ''
+    spec:
+      priorityClassName: system-node-critical
+      affinity:
+        nodeAffinity:
+          requiredDuringSchedulingIgnoredDuringExecution:
+            nodeSelectorTerms:
+            - matchExpressions:
+              - key: "nvidia.com/gpu"
+                operator: Exists
+      tolerations:
+      - operator: "Exists"
+        effect: "NoExecute"
+      - operator: "Exists"
+        effect: "NoSchedule"
+      hostNetwork: true
+      hostPID: true
+      volumes:
+      - name: device-plugin
+        hostPath:
+          path: /var/lib/kubelet/device-plugins
+      - name: dev
+        hostPath:
+          path: /dev
+      containers:
+      - image: "k8s.gcr.io/nvidia-gpu-device-plugin@sha256:0842734032018be107fa2490c98156992911e3e1f2a21e059ff0105b07dd8e9e"
+        command: ["/usr/bin/nvidia-gpu-device-plugin", "-logtostderr"]
+        name: nvidia-gpu-device-plugin
+        resources:
+          requests:
+            cpu: 50m
+            memory: 10Mi
+          limits:
+            cpu: 50m
+            memory: 10Mi
+        securityContext:
+          privileged: true
+        volumeMounts:
+        - name: device-plugin
+          mountPath: /device-plugin
+        - name: dev
+          mountPath: /dev
+  updateStrategy:
+    type: RollingUpdate

roles/kubernetes-apps/container_engine_accelerator/nvidia_gpu/templates/nvidia-driver-install-daemonset.yml.j2

@@ -0,0 +1,80 @@
+# Copyright 2017 Google Inc. All rights reserved.
+#
+# 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.
+
+apiVersion: extensions/v1beta1
+kind: DaemonSet
+metadata:
+  name: nvidia-driver-installer
+  namespace: kube-system
+spec:
+  template:
+    metadata:
+      labels:
+        name: nvidia-driver-installer
+      annotations:
+        scheduler.alpha.kubernetes.io/critical-pod: ''
+    spec:
+      priorityClassName: system-node-critical
+      affinity:
+        nodeAffinity:
+          requiredDuringSchedulingIgnoredDuringExecution:
+            nodeSelectorTerms:
+            - matchExpressions:
+              - key: "nvidia.com/gpu"
+                operator: Exists
+      tolerations:
+      - key: "nvidia.com/gpu"
+        effect: "NoSchedule"
+        operator: "Exists"
+      hostNetwork: true
+      hostPID: true
+      volumes:
+      - name: dev
+        hostPath:
+          path: /dev
+      - name: nvidia-install-dir-host
+        hostPath:
+          path: /home/kubernetes/bin/nvidia
+      - name: root-mount
+        hostPath:
+          path: /
+      initContainers:
+      - image: "{{nvidia_driver_install_container}}"
+        name: nvidia-driver-installer
+        resources:
+          requests:
+            cpu: 0.15
+        securityContext:
+          privileged: true
+        env:
+          - name: NVIDIA_INSTALL_DIR_HOST
+            value: /home/kubernetes/bin/nvidia
+          - name: NVIDIA_INSTALL_DIR_CONTAINER
+            value: /usr/local/nvidia
+          - name: ROOT_MOUNT_DIR
+            value: /root
+          - name: NVIDIA_DRIVER_VERSION
+            value: "{{nvidia_driver_version}}"
+          - name: NVIDIA_DRIVER_DOWNLOAD_URL
+            value: "{{nvidia_driver_download_url_default}}"
+        volumeMounts:
+        - name: nvidia-install-dir-host
+          mountPath: /usr/local/nvidia
+        - name: dev
+          mountPath: /dev
+        - name: root-mount
+          mountPath: /root
+      containers:
+      - image: "gcr.io/google-containers/pause:2.0"
+        name: pause

roles/kubernetes-apps/container_engine_accelerator/nvidia_gpu/vars/centos-7.yml

@@ -0,0 +1,3 @@
+---
+nvidia_driver_install_container: atzedevries/nvidia-centos-driver-installer:2
+nvidia_driver_install_supported: true

roles/kubernetes-apps/container_engine_accelerator/nvidia_gpu/vars/ubuntu-16.yml

@@ -0,0 +1,3 @@
+---
+nvidia_driver_install_container: gcr.io/google-containers/ubuntu-nvidia-driver-installer@sha256:eea7309dc4fa4a5c9d716157e74b90826e0a853aa26c7219db4710ddcd1ad8bc
+nvidia_driver_install_supported: true

roles/kubernetes-apps/container_engine_accelerator/nvidia_gpu/vars/ubuntu-18.yml

@@ -0,0 +1,3 @@
+---
+nvidia_driver_install_container: gcr.io/google-containers/ubuntu-nvidia-driver-installer@sha256:eea7309dc4fa4a5c9d716157e74b90826e0a853aa26c7219db4710ddcd1ad8bc
+nvidia_driver_install_supported: true

roles/kubernetes-apps/meta/main.yml

@@ -37,6 +37,12 @@ dependencies:
       - apps
       - persistent_volumes
 
+  - role: kubernetes-apps/container_engine_accelerator
+    when: nvidia_accelerator_enabled
+    tags:
+      - apps
+      - container_engine_accelerator
+
   - role: kubernetes-apps/cloud_controller/oci
     when:
       - cloud_provider is defined

roles/kubernetes/node/templates/kubelet.standard.env.j2

@@ -100,6 +100,11 @@ KUBELET_HOSTNAME="--hostname-override={{ kube_override_hostname }}"
 {% else %}
 {%   set dummy = role_node_labels.append('node-role.kubernetes.io/node=true') %}
 {% endif %}
+{% if nvidia_gpu_nodes is defined and nvidia_accelerator_enabled|bool %}
+{%   if inventory_hostname in nvidia_gpu_nodes %}
+{%     set dummy = role_node_labels.append('nvidia.com/gpu=true')  %}
+{%   endif %}
+{% endif %}
 {% set inventory_node_labels = [] %}
 {% if node_labels is defined %}
 {%   for labelname, labelvalue in node_labels.iteritems() %}
@@ -108,7 +113,15 @@ KUBELET_HOSTNAME="--hostname-override={{ kube_override_hostname }}"
 {% endif %}
 {% set all_node_labels = role_node_labels + inventory_node_labels %}
 
+{# Kubelet node taints for gpu #}
+{% if nvidia_gpu_nodes is defined and nvidia_accelerator_enabled|bool %}
+{%   if inventory_hostname in nvidia_gpu_nodes %}
+{%     set kubelet_args_kubeconfig %}{{ kubelet_args_kubeconfig }} --register-with-taints=nvidia.com/gpu=:NoSchedule{% endset %}
+{%   endif %}
+{% endif %}
+
 KUBELET_ARGS="{{ kubelet_args_base }} {{ kubelet_args_dns }} {{ kubelet_args_kubeconfig }} {{ kube_reserved }} --node-labels={{ all_node_labels | join(',') }} {% if kube_feature_gates %} --feature-gates={{ kube_feature_gates|join(',') }} {% endif %} {% if kubelet_custom_flags is string %} {{kubelet_custom_flags}} {% else %}{% for flag in kubelet_custom_flags %} {{flag}} {% endfor %}{% endif %}"
+
 {% if kube_network_plugin is defined and kube_network_plugin in ["calico", "canal", "flannel", "weave", "contiv", "cilium"] %}
 KUBELET_NETWORK_PLUGIN="--network-plugin=cni --cni-conf-dir=/etc/cni/net.d --cni-bin-dir=/opt/cni/bin"
 {% elif kube_network_plugin is defined and kube_network_plugin == "weave" %}