大网数据广告活动
Kubernetes 部署与运维

Install Drivers and Allocate Devices with DRA

· 52CE 我爱测

drivers in your cluster and how to use them in conjunction with the DRA APIs to allocate to Pods. This page is intended for cluster administrators.

This tutorial shows you how to install

drivers in your cluster and how to use them in conjunction with the DRA APIs to allocate to Pods. This page is intended for cluster administrators.

lets a cluster manage availability and allocation of hardware resources to satisfy Pod-based claims for hardware requirements and preferences. To support this, a mixture of Kubernetes built-in components (like the Kubernetes scheduler, kubelet, and kube-controller-manager) and third-party drivers from device owners (called DRA drivers) share the responsibility to advertise, allocate, prepare, mount, healthcheck, unprepare, and cleanup resources throughout the Pod lifecycle. These components share information via a series of DRA specific APIs in the resource.k8s.io API group including , , , as well as new fields in the Pod spec itself.

  • Deploy an example DRA driver
  • Deploy a Pod requesting a hardware claim using DRA APIs
  • Delete a Pod that has a claim

Your cluster should support RBAC. You can try this tutorial with a cluster using a different authorization mechanism, but in that case you will have to adapt the steps around defining roles and permissions.

This tutorial has been tested with Linux nodes, though it may also work with other types of nodes.

If your cluster is not currently running Kubernetes then please check the documentation for the version of Kubernetes that you plan to use.

Explore the initial cluster state {#explore-initial-state}

You can spend some time to observe the initial state of a cluster with DRA enabled, especially if you have not used these APIs extensively before. If you set up a new cluster for this tutorial, with no driver installed and no Pod claims yet to satisfy, the output of these commands won't show any resources.

  1. Get a list of :

    kubectl get deviceclasses
    

    The output is similar to this:

    No resources found
    
  2. Get a list of :

    kubectl get resourceslices
    

    The output is similar to this:

    No resources found
    
  3. Get a list of and .

    kubectl get resourceclaims -A
    kubectl get resourceclaimtemplates -A
    

    The output is similar to this:

    No resources found
    No resources found
    

At this point, you have confirmed that DRA is enabled and configured properly in the cluster, and that no DRA drivers have advertised any resources to the DRA APIs yet.

Install an example DRA driver {#install-example-driver}

DRA drivers are third-party applications that run on each node of your cluster to interface with the hardware of that node and Kubernetes' built-in DRA components. The installation procedure depends on the driver you choose, but is likely deployed as a to all or a selection of the nodes (using or similar mechanisms) in your cluster.

Check your driver's documentation for specific installation instructions, which might include a Helm chart, a set of manifests, or other deployment tooling.

This tutorial uses an example driver which can be found in the kubernetes-sigs/dra-example-driver repository to demonstrate driver installation. This example driver advertises simulated GPUs to Kubernetes for your Pods to interact with.

Prepare your cluster for driver installation {#prepare-cluster-driver}

To simplify cleanup, create a namespace named dra-tutorial:

  1. Create the namespace:

    kubectl create namespace dra-tutorial 
    

    In a production environment, you would likely be using a previously released or qualified image from the driver vendor or your own organization, and your nodes would need to have access to the image registry where the driver image is hosted. In this tutorial, you will use a publicly released image of the dra-example-driver to simulate access to a DRA driver image.

  2. Confirm your nodes have access to the image by running the following from within one of your cluster's nodes:

    docker pull registry.k8s.io/dra-example-driver/dra-example-driver:v0.2.0
    

Deploy the DRA driver components

For this tutorial, you will install the critical example resource driver components individually with kubectl.

  1. Create the DeviceClass representing the device types this DRA driver supports:

    kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/deviceclass.yaml
    
  2. Create the ServiceAccount, ClusterRole and ClusterRoleBinding that will be used by the driver to gain permissions to interact with the Kubernetes API on this cluster:

    1. Create the Service Account:

      kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/serviceaccount.yaml
      
    2. Create the ClusterRole:

      kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/clusterrole.yaml
      
    3. Create the ClusterRoleBinding:

      kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/clusterrolebinding.yaml
      
  3. Create a for the DRA driver. The PriorityClass prevents preemption of th DRA driver component, which is responsible for important lifecycle operations for Pods with claims. Learn more about pod priority and preemption here.

    kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/priorityclass.yaml
    
  4. Deploy the actual DRA driver as a DaemonSet configured to run the example driver binary with the permissions provisioned above. The DaemonSet has the permissions that you granted to the ServiceAccount in the previous steps.

    kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/daemonset.yaml
    

    The DaemonSet is configured with the volume mounts necessary to interact with the underlying Container Device Interface (CDI) directory, and to expose its socket to kubelet via the kubelet/plugins directory.

Verify the DRA driver installation {#verify-driver-install}

  1. Get a list of the Pods of the DRA driver DaemonSet across all worker nodes:

    kubectl get pod -l app.kubernetes.io/name=dra-example-driver -n dra-tutorial
    

    The output is similar to this:

    NAME                                     READY   STATUS    RESTARTS   AGE
    dra-example-driver-kubeletplugin-4sk2x   1/1     Running   0          13s
    dra-example-driver-kubeletplugin-cttr2   1/1     Running   0          13s
    
  2. The initial responsibility of each node's local DRA driver is to update the cluster with what devices are available to Pods on that node, by publishing its metadata to the ResourceSlices API. You can check that API to see that each node with a driver is advertising the device class it represents.

    Check for available ResourceSlices:

    kubectl get resourceslices
    

    The output is similar to this:

    NAME                                 NODE           DRIVER            POOL           AGE
    kind-worker-gpu.example.com-k69gd    kind-worker    gpu.example.com   kind-worker    19s
    kind-worker2-gpu.example.com-qdgpn   kind-worker2   gpu.example.com   kind-worker2   19s
    

At this point, you have successfully installed the example DRA driver, and confirmed its initial configuration. You're now ready to use DRA to schedule Pods.

Claim resources and deploy a Pod {#claim-resources-pod}

To request resources using DRA, you create ResourceClaims or ResourceClaimTemplates that define the resources that your Pods need. In the example driver, a memory capacity attribute is exposed for mock GPU devices. This section shows you how to use to express your requirements in a ResourceClaim, select that ResourceClaim in a Pod specification, and observe the resource allocation.

This tutorial showcases only one basic example of a DRA ResourceClaim. Read Dynamic Resource Allocation to learn more about ResourceClaims.

Create the ResourceClaim

In this section, you create a ResourceClaim and reference it in a Pod. Whatever the claim, the deviceClassName is a required field, narrowing down the scope of the request to a specific device class. The request itself can include a expression that references attributes that may be advertised by the driver managing that device class.

In this example, you will create a request for any GPU advertising over 10Gi memory capacity. The attribute exposing capacity from the example driver takes the form device.capacity['gpu.example.com'].memory. Note also that the name of the claim is set to some-gpu.

kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/example/resourceclaim.yaml

Create the Pod that references that ResourceClaim

Below is the Pod manifest referencing the ResourceClaim you just made, some-gpu, in the spec.resourceClaims.resourceClaimName field. The local name for that claim, gpu, is then used in the spec.containers.resources.claims.name field to allocate the claim to the Pod's underlying container.

kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/example/pod.yaml

Confirm the pod has deployed:

kubectl get pod pod0 -n dra-tutorial

The output is similar to this:

NAME   READY   STATUS    RESTARTS   AGE
pod0   1/1     Running   0          9s

Explore the DRA state

After you create the Pod, the cluster tries to schedule that Pod to a node where Kubernetes can satisfy the ResourceClaim. In this tutorial, the DRA driver is deployed on all nodes, and is advertising mock GPUs on all nodes, all of which have enough capacity advertised to satisfy the Pod's claim, so Kubernetes can schedule this Pod on any node and can allocate any of the mock GPUs on that node.

When Kubernetes allocates a mock GPU to a Pod, the example driver adds environment variables in each container it is allocated to in order to indicate which GPUs would have been injected into them by a real resource driver and how they would have been configured, so you can check those environment variables to see how the Pods have been handled by the system.

  1. Check the Pod logs, which report the name of the mock GPU that was allocated:

    kubectl logs pod0 -c ctr0 -n dra-tutorial | grep -E "GPU_DEVICE_[0-9]+=" | grep -v "RESOURCE_CLAIM"
    

    The output is similar to this:

    declare -x GPU_DEVICE_0="gpu-0"
    
  2. Check the state of the ResourceClaim object:

    kubectl get resourceclaims -n dra-tutorial
    

    The output is similar to this:

    NAME       STATE                AGE
    some-gpu   allocated,reserved   34s
    

    In this output, the STATE column shows that the ResourceClaim is allocated and reserved.

  3. Check the details of the some-gpu ResourceClaim. The status stanza of the ResourceClaim has information about the allocated device and the Pod it has been reserved for:

    kubectl get resourceclaim some-gpu -n dra-tutorial -o yaml
    

    The output is similar to this:

    apiVersion: resource.k8s.io/v1 kind: ResourceClaim metadata: creationTimestamp: "2025-08-20T18:17:31Z" finalizers:

    • resource.kubernetes.io/delete-protection name: some-gpu namespace: dra-tutorial resourceVersion: "2326" uid: d3e48dbf-40da-47c3-a7b9-f7d54d1051c3 spec: devices: requests:
      • exactly: allocationMode: ExactCount count: 1 deviceClassName: gpu.example.com selectors:
        • cel: expression: device.capacity['gpu.example.com'].memory.compareTo(quantity('10Gi'))

          = 0 name: some-gpu status: allocation: devices: results:

        • device: gpu-0 driver: gpu.example.com pool: kind-worker request: some-gpu nodeSelector: nodeSelectorTerms:
        • matchFields:
          • key: metadata.name operator: In values:
            • kind-worker reservedFor:
    • name: pod0 resource: pods uid: c4dadf20-392a-474d-a47b-ab82080c8bd7
  4. To check how the driver handled device allocation, get the logs for the driver DaemonSet Pods:

    kubectl logs -l app.kubernetes.io/name=dra-example-driver -n dra-tutorial
    

    The output is similar to this:

    I0820 18:17:44.131324       1 driver.go:106] PrepareResourceClaims is called: number of claims: 1
    I0820 18:17:44.135056       1 driver.go:133] Returning newly prepared devices for claim 'd3e48dbf-40da-47c3-a7b9-f7d54d1051c3': [{[some-gpu] kind-worker gpu-0 [k8s.gpu.example.com/gpu=common k8s.gpu.example.com/gpu=d3e48dbf-40da-47c3-a7b9-f7d54d1051c3-gpu-0]}]
    

You have now successfully deployed a Pod that claims devices using DRA, verified that the Pod was scheduled to an appropriate node, and saw that the associated DRA APIs kinds were updated with the allocation status.

Delete a Pod that has a claim {#delete-pod-claim}

When a Pod with a claim is deleted, the DRA driver deallocates the resource so it can be available for future scheduling. To validate this behavior, delete the Pod that you created in the previous steps and watch the corresponding changes to the ResourceClaim and driver.

Delete the pod0 Pod:

kubectl delete pod pod0 -n dra-tutorial

The output is similar to this:

pod "pod0" deleted

Observe the DRA state

When the Pod is deleted, the driver deallocates the device from the ResourceClaim and updates the ResourceClaim resource in the Kubernetes API. The ResourceClaim has a pending state until it's referenced in a new Pod.

  1. Check the state of the some-gpu ResourceClaim:

    kubectl get resourceclaims -n dra-tutorial
    

    The output is similar to this:

    NAME       STATE     AGE
    some-gpu   pending   76s
    
  2. Verify that the driver has processed unpreparing the device for this claim by checking the driver logs:

    kubectl logs -l app.kubernetes.io/name=dra-example-driver -n dra-tutorial
    

    The output is similar to this:

    I0820 18:22:15.629376       1 driver.go:138] UnprepareResourceClaims is called: number of claims: 1
    

You have now deleted a Pod that had a claim, and observed that the driver took action to unprepare the underlying hardware resource and update the DRA APIs to reflect that the resource is available again for future scheduling.

To clean up the resources that you created in this tutorial, follow these steps:

kubectl delete namespace dra-tutorial
kubectl delete deviceclass gpu.example.com
kubectl delete clusterrole dra-example-driver-role
kubectl delete clusterrolebinding dra-example-driver-role-binding
kubectl delete priorityclass dra-driver-high-priority

来源:Kubernetes 文档贡献者 · 作者与贡献记录 · CC BY 4.0。离线整理了版式、链接和媒体显示。

资料来源与适用范围

资料来源
Kubernetes 文档贡献者
适用范围
请按文中版本与环境说明核对
内容语言
中文;命令、路径和错误标识保留技术原文
本地收录日期
2026-11-15
授权说明
CC BY 4.0

正文与所需资源由本站本地存储。查看来源及更新记录 ↗

请按适用版本和实际日志确认原因;本站没有对每一种系统或设备逐项执行文中的操作。