starlingx/docs
Code Issues Proposed changes

Merge "Nvidia gpu-operator deployment updates"

Browse Source
This commit is contained in:
Zuul
2021-12-06 17:14:33 +00:00
committed by Gerrit Code Review
parent 6a28427f4b 0c13a27861
commit acdf534b63
2 changed files with 21 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

50
doc/source/node_management/kubernetes/hardware_acceleration_devices/configure-nvidia-gpu-operator-for-pci-passthrough.rst
View File

@@ -39,9 +39,9 @@ software components. These components include:
.. rubric:: |prereq| .. rubric:: |prereq|
Download the **gpu-operator-v3-1.6.0.3.tgz** file at Download the **gpu-operator-v3-1.8.1.4.tgz** file at
`http://mirror.starlingx.cengn.ca/mirror/starlingx/ `http://mirror.starlingx.cengn.ca/mirror/starlingx/release/latest_release/centos/containers/inputs/downloads/
<http://mirror.starlingx.cengn.ca/mirror/starlingx/>`__. <http://mirror.starlingx.cengn.ca/mirror/starlingx/release/latest_release/centos/containers/inputs/downloads/>`__.
Use the following steps to configure the GPU Operator container: Use the following steps to configure the GPU Operator container:
@@ -53,7 +53,8 @@ Use the following steps to configure the GPU Operator container:
~(keystone_admin)]$ system host-lock <hostname> ~(keystone_admin)]$ system host-lock <hostname>
#. Configure the Container Runtime host path to the NVIDIA runtime which will be installed by the GPU Operator Helm deployment. #. Configure the Container Runtime host path to the NVIDIA runtime which will
be installed by the GPU Operator Helm deployment.
.. code-block:: none .. code-block:: none
@@ -65,41 +66,32 @@ Use the following steps to configure the GPU Operator container:
~(keystone_admin)]$ system host-unlock <hostname> ~(keystone_admin)]$ system host-unlock <hostname>
#. Create the RuntimeClass resource definition and apply it to the system.
.. code-block:: none
cat > nvidia.yml << EOF
kind: RuntimeClass
apiVersion: node.k8s.io/v1beta1
metadata:
name: nvidia
handler: nvidia
EOF
.. code-block:: none
~(keystone_admin)]$ kubectl apply -f nvidia.yml
#. Install the GPU Operator Helm charts. #. Install the GPU Operator Helm charts.
.. code-block:: none .. code-block:: none
~(keystone_admin)]$ helm install gpu-operator /path/to/gpu-operator-1.6.0.3.tgz ~(keystone_admin)]$ helm install gpu-operator /path/to/gpu-operator-v3-1.8.1.4.tgz
#. Check if the GPU Operator is deployed using the following command. #. Check if the GPU Operator is deployed using the following command.
.. code-block:: none .. code-block:: none
~(keystone_admin)]$ kubectl get pods –A ~(keystone_admin)]$ kubectl get pods –A
NAMESPACE NAME READY STATUS RESTART AGE NAMESPACE NAME READY STATUS RESTARTS AGE
default gpu-operator-596c49cb9b-2tdlw 1/1 Running 1 24h ..............
default gpu-operator-node-feature-discovery-master-7f87b4d6bb-wsbn4 1/1 Running 2 24h default gpu-operator-5dddfcbb58-xpwbh 1/1 Running 0 3m13s
default gpu-operator-node-feature-discovery-worker-hqzvw 1/1 Running 4 24h default gpu-operator-node-feature-discovery-master-58d884d5cc-56qch 1/1 Running 0 3m13s
gpu-operator-resources nvidia-container-toolkit-daemonset-8f7nl 1/1 Running 0 14h default gpu-operator-node-feature-discovery-worker-p495j 1/1 Running 0 3m13s
gpu-operator-resources nvidia-device-plugin-daemonset-g9lmk 1/1 Running 0 14h gpu-operator-resources gpu-feature-discovery-swmj8 1/1 Running 0 2m52s
gpu-operator-resources nvidia-device-plugin-validation 0/1 Pending 0 24h gpu-operator-resources nvidia-cuda-validator-zcfp9 0/1 Completed 0 2m31s
gpu-operator-resources nvidia-driver-daemonset-9mnwr 1/1 Running 0 14h gpu-operator-resources nvidia-dcgm-9447k 1/1 Running 0 2m52s
gpu-operator-resources nvidia-dcgm-exporter-9c82q 1/1 Running 0 2m52s
gpu-operator-resources nvidia-device-plugin-daemonset-ljm4q 1/1 Running 0 2m52s
gpu-operator-resources nvidia-device-plugin-validator-j9kjz 0/1 Completed 0 2m25s
gpu-operator-resources nvidia-driver-daemonset-qph2s 1/1 Running 0 2m52s
gpu-operator-resources nvidia-operator-validator-dw6sc 1/1 Running 0 2m52s
..........
kube-system toolkit-installer-xzrt8 1/1 Running 0 3m13s
The plugin validation pod is marked completed. The plugin validation pod is marked completed.

1
doc/source/node_management/kubernetes/hardware_acceleration_devices/delete-the-gpu-operator.rst
View File

@@ -21,7 +21,6 @@ Ensure that all user generated pods with access to `nvidia.com/gpu` resources ar
.. code-block:: none .. code-block:: none
~(keystone_admin)]$ helm delete gpu-operator ~(keystone_admin)]$ helm delete gpu-operator
~(keystone_admin)]$ kubectl delete runtimeclasses.node.k8s.io nvidia
#. Remove the GPU Operator, and remove the service parameter platform #. Remove the GPU Operator, and remove the service parameter platform
`container\_runtime custom\_container\_runtime` from the system, using the `container\_runtime custom\_container\_runtime` from the system, using the