github
/
k3s
mirror of https://github.com/k3s-io/k3s
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
k3s/cluster/addons/device-plugins/nvidia-gpu/daemonset.yaml

60 lines
1.6 KiB
YAML
Raw Normal View History

Update nvidia-gpu-device-plugin to apps/v1 and use RollingUpdate updateStrategy. Even though RollingUpdate is the default updateStrategy, we need to specify it explicitly here because otherwise updating from extensions/v1beta1 to apps/v1 doesn't change the updateStrategy.
2018-05-25 00:53:07 +00:00
apiVersion: apps/v1
Run nvidia-gpu device-plugin daemonset as an addon on GCE nodes that have nvidia GPUs attached.
2017-10-30 22:50:08 +00:00
kind: DaemonSet
metadata:
name: nvidia-gpu-device-plugin
namespace: kube-system
labels:
k8s-app: nvidia-gpu-device-plugin
addonmanager.kubernetes.io/mode: Reconcile
spec:
Update nvidia-gpu-device-plugin to apps/v1 and use RollingUpdate updateStrategy. Even though RollingUpdate is the default updateStrategy, we need to specify it explicitly here because otherwise updating from extensions/v1beta1 to apps/v1 doesn't change the updateStrategy.
2018-05-25 00:53:07 +00:00
selector:
matchLabels:
k8s-app: nvidia-gpu-device-plugin
Run nvidia-gpu device-plugin daemonset as an addon on GCE nodes that have nvidia GPUs attached.
2017-10-30 22:50:08 +00:00
template:
metadata:
labels:
k8s-app: nvidia-gpu-device-plugin
Changes nvidia-gpu device plugin addon config settings: - Runs as system critical pod - Makes resource limits to match its resource requets - Modifies test/e2e/scheduling/nvidia-gpus.go to cope with the recent change of running the device plugin as a system addon. - The resource settings of the addon is based on the test results from 8 nvidia-tesla-k80 gpus.
2017-11-20 18:55:10 +00:00
annotations:
scheduler.alpha.kubernetes.io/critical-pod: ''
Run nvidia-gpu device-plugin daemonset as an addon on GCE nodes that have nvidia GPUs attached.
2017-10-30 22:50:08 +00:00
spec:
Changes nvidia-gpu device plugin addon config settings: - Runs as system critical pod - Makes resource limits to match its resource requets - Modifies test/e2e/scheduling/nvidia-gpus.go to cope with the recent change of running the device plugin as a system addon. - The resource settings of the addon is based on the test results from 8 nvidia-tesla-k80 gpus.
2017-11-20 18:55:10 +00:00
priorityClassName: system-node-critical
Run nvidia-gpu device-plugin daemonset as an addon on GCE nodes that have nvidia GPUs attached.
2017-10-30 22:50:08 +00:00
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: cloud.google.com/gke-accelerator
operator: Exists
nvidia-gpu-device-plugin daemonset should tolerate nvidia.com/gpu taint. It is expected that nodes with extended resources attached will be tainted with the resouce name, so that we can create dedicated nodes. If ExtendedResourceToleration admission controller is enabled, pods requesting such resources will automatically tolerate such taints. nvidia-gpu-device-plugin daemonset doesn't request such resources but still needs to run on such nodes, so it needs this toleration.
2017-11-29 19:31:39 +00:00
tolerations:
Add wildcard toleration to nvidia-gpu-device-plugin. This addon runs only on nodes with GPUs and it's essential for their proper function.
2018-04-06 20:53:40 +00:00
- operator: "Exists"
effect: "NoExecute"
- operator: "Exists"
nvidia-gpu-device-plugin daemonset should tolerate nvidia.com/gpu taint. It is expected that nodes with extended resources attached will be tainted with the resouce name, so that we can create dedicated nodes. If ExtendedResourceToleration admission controller is enabled, pods requesting such resources will automatically tolerate such taints. nvidia-gpu-device-plugin daemonset doesn't request such resources but still needs to run on such nodes, so it needs this toleration.
2017-11-29 19:31:39 +00:00
effect: "NoSchedule"
Run nvidia-gpu device-plugin daemonset as an addon on GCE nodes that have nvidia GPUs attached.
2017-10-30 22:50:08 +00:00
volumes:
- name: device-plugin
hostPath:
path: /var/lib/kubelet/device-plugins
- name: dev
hostPath:
path: /dev
containers:
Update gke nvidia-gpu-device-plugin to the latest version that supports both v1alpha and v1beta1 device plugin versions. Re-enables nvidia-gpus e2e test after verifying the test passes now.
2018-02-23 01:27:09 +00:00
- image: "k8s.gcr.io/nvidia-gpu-device-plugin@sha256:0842734032018be107fa2490c98156992911e3e1f2a21e059ff0105b07dd8e9e"
Run nvidia-gpu device-plugin daemonset as an addon on GCE nodes that have nvidia GPUs attached.
2017-10-30 22:50:08 +00:00
command: ["/usr/bin/nvidia-gpu-device-plugin", "-logtostderr"]
name: nvidia-gpu-device-plugin
resources:
requests:
Changes nvidia-gpu device plugin addon config settings: - Runs as system critical pod - Makes resource limits to match its resource requets - Modifies test/e2e/scheduling/nvidia-gpus.go to cope with the recent change of running the device plugin as a system addon. - The resource settings of the addon is based on the test results from 8 nvidia-tesla-k80 gpus.
2017-11-20 18:55:10 +00:00
cpu: 50m
memory: 10Mi
limits:
cpu: 50m
Run nvidia-gpu device-plugin daemonset as an addon on GCE nodes that have nvidia GPUs attached.
2017-10-30 22:50:08 +00:00
memory: 10Mi
securityContext:
privileged: true
volumeMounts:
- name: device-plugin
mountPath: /device-plugin
- name: dev
mountPath: /dev
Update nvidia-gpu-device-plugin to apps/v1 and use RollingUpdate updateStrategy. Even though RollingUpdate is the default updateStrategy, we need to specify it explicitly here because otherwise updating from extensions/v1beta1 to apps/v1 doesn't change the updateStrategy.
2018-05-25 00:53:07 +00:00
updateStrategy:
type: RollingUpdate