diff --git a/README.md b/README.md index 0697f4e6a..0ce45e5dd 100644 --- a/README.md +++ b/README.md @@ -2,6 +2,10 @@ [![End-to-end Tests](https://github.com/NVIDIA/k8s-device-plugin/actions/workflows/e2e.yaml/badge.svg)](https://github.com/NVIDIA/k8s-device-plugin/actions/workflows/e2e.yaml) [![Go Report Card](https://goreportcard.com/badge/github.com/NVIDIA/k8s-device-plugin)](https://goreportcard.com/report/github.com/NVIDIA/k8s-device-plugin) [![Latest Release](https://img.shields.io/github/v/release/NVIDIA/k8s-device-plugin)](https://github.com/NVIDIA/k8s-device-plugin/releases/latest) +> This fork of the NVIDIA device plugin for Kubernetes has a `release-1.1` branch that is based on the `v0.16.1` tag of the original NVIDIA repository. This version includes all the features and updates available in the `v0.16.1` tag, along with any additional modifications specific to this fork. +> +> For more details on the changes and updates in this release, please refer to the [Documentation](docs/resource-naming/README.md). + ## Table of Contents - [About](#about) diff --git a/cmd/nvidia-device-plugin/main.go b/cmd/nvidia-device-plugin/main.go index 7424c7c9c..68a3374bf 100644 --- a/cmd/nvidia-device-plugin/main.go +++ b/cmd/nvidia-device-plugin/main.go @@ -34,7 +34,6 @@ import ( spec "github.com/NVIDIA/k8s-device-plugin/api/config/v1" "github.com/NVIDIA/k8s-device-plugin/internal/info" - "github.com/NVIDIA/k8s-device-plugin/internal/logger" "github.com/NVIDIA/k8s-device-plugin/internal/plugin" "github.com/NVIDIA/k8s-device-plugin/internal/rm" "github.com/NVIDIA/k8s-device-plugin/internal/watch" @@ -280,7 +279,14 @@ func startPlugins(c *cli.Context, flags []cli.Flag) ([]plugin.Interface, bool, e if err != nil { return nil, false, fmt.Errorf("unable to load config: %v", err) } - spec.DisableResourceNamingInConfig(logger.ToKlog, config) + + // This block has been commented out due to issue #69. + // Date: 2024-08-07 + // Reason: Commenting out this block allows for the configuration of resource naming. + // This enables the setting of different quotas for various GPU types. + // For more details, see the GitHub issue: https://github.com/volcano-sh/devices/issues/69 + + // spec.DisableResourceNamingInConfig(logger.ToKlog, config) driverRoot := root(*config.Flags.Plugin.ContainerDriverRoot) // We construct an NVML library specifying the path to libnvidia-ml.so.1 diff --git a/docs/resource-naming/README.md b/docs/resource-naming/README.md new file mode 100644 index 000000000..90cc6952b --- /dev/null +++ b/docs/resource-naming/README.md @@ -0,0 +1,225 @@ + +# Configuring the NVIDIA Device Plugin to report Different GPU Models with Volcano Capacity Scheduling + +Volcano v1.9.0 introduces capacity scheduling capabilities, allowing you to configure different quotas for various GPU types (essential in production environments). For example: + +```yaml +apiVersion: scheduling.volcano.sh/v1beta1 +kind: Queue +metadata: + name: queue1 +spec: + reclaimable: true + deserved: # set the deserved field. + cpu: 2 + memory: 8Gi + nvidia.com/t4: 40 + nvidia.com/a100: 20 +``` + +However, the default Nvidia Device Plugin reports GPU resources as `nvidia.com/gpu`, which does not support reporting different GPU models as shown in the example. + +This guide walks you through configuring the NVIDIA Device Plugin to report different GPU models and integrating it with Volcano's capacity scheduling. + +## Install a Customized Device Plugin + +In this section, we will cover how to install a customized Device Plugin using Helm. Instructions for installing `helm` can be found [here](https://helm.sh/docs/intro/install/). For prerequisites to install the Device Plugin, refer to the [documentation](../../README.md#prerequisites) for detailed instructions. + +If you are using the GPU Operator to manage all GPU-related components, there are two options. One option is to disable the GPU Operator's management of the Device Plugin and follow the steps in this section. The other option is to modify the GPU Operator configuration accordingly. + +### 1.1 Install a Custom Device Plugin + +Begin by setting up the plugin's `helm` repository and updating it as follows: + +```shell +helm repo add nvdp https://nvidia.github.io/k8s-device-plugin +helm repo update +``` + +Then verify if the version v0.16.1 on which the modified release of the plugin is based exists: + +```shell +$ helm search repo nvdp --version 0.16.1 +NAME CHART VERSION APP VERSION DESCRIPTION +nvdp/nvidia-device-plugin 0.16.1 0.16.1 A Helm chart for ... +``` + +Next, prepare a ConfigMap file, assuming it's named `config.yaml`. A typical configuration is as follows: + +```yaml +version: v1 +flags: + migStrategy: "none" + failOnInitError: true + nvidiaDriverRoot: "/" + plugin: + passDeviceSpecs: false + deviceListStrategy: "envvar" + deviceIDStrategy: "uuid" +``` + +Once this repo is updated and the ConfigMap is prepared, you can begin installing packages from it to deploy the `volcano-device-plugin`. + +```shell +helm upgrade -i nvdp nvdp/nvidia-device-plugin \ + --version=0.16.1 \ + --namespace nvidia-device-plugin \ + --create-namespace \ + --set gfd.enabled=true \ + --set image.repository=volcanosh/volcano-device-plugin \ + --set image.tag=v1.1.0 \ + --set config.default=default-config \ + --set-file config.map.default-config=config.yaml +``` + +Here is a brief explanation of some of the configuration parameters: + +| Command | Usage | +| ------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `--set gfd.enabled=true` | Enables the GFD feature. If you already have NFD deployed on your cluster and do not wish for it to be pulled in by this installation, you can disable it with `nfd.enabled=false`. | +| `--set image.repository`
`--set image.tag` | Replaces the official image with the modified version. | +| `--set config.default`
`--set-file config.map` | Sets the ConfigMap for the Device Plugin to configure resource renaming rules. | + +If you wish to set multiple ConfigMaps to allow different policies for different nodes or configure Time Slicing, refer to the [documentation](../../README.md#multiple-config-file-example) for more information. + +The _Volcano_ community provides an image based on NVIDIA Device Plugin 0.16.1. If this version does not meet your needs, you can refer to the [discussion](https://github.com/NVIDIA/k8s-device-plugin/issues/424) and create your custom image that supports Resource Naming. + +### 1.2 Set Resource Renaming Rules + +Once the Device Plugin is deployed, to report resources like `nvidia.com/a100`, you need to update the ConfigMap to provide renaming rules, for example: + +```yaml +version: v1 +flags: + migStrategy: "none" + failOnInitError: true + nvidiaDriverRoot: "/" + plugin: + passDeviceSpecs: false + deviceListStrategy: envvar + deviceIDStrategy: uuid +# new +resources: + gpus: + - pattern: "A100-*-40GB" + name: a100-40gb + - pattern: "*A100*" + name: a100 +``` + +The `pattern` field can be checked on your node by running the command: + +```shell +$ nvidia-smi --query-gpu=name --format=csv,noheader +A100-SXM4-40GB +... +``` + +Patterns can include wildcards (i.e. ‘*’) to match against multiple devices with similar names. Additionally, the order of the entries under `resources.gpus` and `resources.mig` matters. Entries earlier in the list will be matched before entries later in the list. MIG devices can also be renamed. For more information, you can read the [documentation](https://docs.google.com/document/d/1dL67t9IqKC2-xqonMi6DV7W2YNZdkmfX7ibB6Jb-qmk). + +After modifying the ConfigMap, you should observe that the node is now reporting `nvidia.com/a100` resources, indicating that the configuration was successful. + +### 1.3 Clean Up Stale Resources + +If you previously installed the Device Plugin and reported `nvidia.com/gpu` type resources, you might notice that `nvidia.com/gpu` is still retained after reconfiguring the Device Plugin. For a detailed discussion on this issue, refer to [here](https://github.com/NVIDIA/k8s-device-plugin/issues/240). + +```yaml +allocatable: + nvidia.com/gpu: 0 + nvidia.com/a100: 8 +``` + + To clean up stale resources, you can start `kubectl proxy` in one terminal: + +```shell +$ kubectl proxy +Starting to serve on 127.0.0.1:8001 +``` + +And in another terminal, run the cleanup script (note `/` needs to be escaped as `~1`): + +```bash +#!/bin/bash + +# Check if at least one node name is provided +if [ "$#" -lt 1 ]; then + echo "Usage: $0 [...]" + exit 1 +fi + +# Prepare the JSON patch data +PATCH_DATA=$(cat <- + nvidia.com/a100,nvidia.com/h100 +``` + +## 3. Configure Volcano to Use the Capacity Scheduling Plugin + +After completing the above configuration, you can edit the `volcano-scheduler-configmap` to enable the capacity plugin: + +```yaml +kind: ConfigMap +apiVersion: v1 +metadata: + name: volcano-scheduler-configmap + namespace: volcano-system +data: + volcano-scheduler.conf: | + actions: "enqueue, allocate, backfill, reclaim" + tiers: + - plugins: + - name: priority + - name: gang + enablePreemptable: false + - name: conformance + - plugins: + - name: drf + enablePreemptable: false + - name: predicates + - name: capacity # add this field and remove proportion plugin. + - name: nodeorder + - name: binpack +``` + +You can customize your configuration according to your needs. For more information, refer to the [documentation](https://github.com/volcano-sh/volcano/blob/release-1.10/docs/user-guide/how_to_use_capacity_plugin.md). \ No newline at end of file