Install Drivers and Allocate Devices with DRA
Kubernetes Docs · نسخهٔ f284a9e7c87f · CC BY 4.0 برای مستندات و مجوزهای همراه
<!-- FUTURE MAINTAINERS: The original point of this doc was for people (mainly cluster administrators) to understand the importance of the DRA driver and its interaction with the DRA APIs. As a result it was a requirement of this tutorial to not use Helm and to be more direct with all the component installation procedures. While much of this content is also useful to workload authors, I see the primary audience of _this_ tutorial as cluster administrators, who I feel also need to understand how the DRA APIs interact with the driver to maintain them well. If I had to choose which audience to focus on in this doc, I would choose cluster administrators. If the prose gets too muddied by considering both of them, I'd rather make a second tutorial for the workload authors that doesn't go into the driver at all (since IMHO that is more representative of what we think their experience should be like) and also potentially goes into much more detailed/ ✨ fun ✨ use cases. -->
{{< feature-state feature_gate_name="DynamicResourceAllocation" >}}
<!-- overview --> This tutorial shows you how to install {{< glossary_tooltip term_id="dra" text="Dynamic Resource Allocation (DRA)" >}} drivers in your cluster and how to use them in conjunction with the DRA APIs to allocate {{< glossary_tooltip text="devices" term_id="device" >}} to Pods. This page is intended for cluster administrators.
{{< glossary_tooltip text="Dynamic Resource Allocation (DRA)" term_id="dra" >}} 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 {{< glossary_tooltip text="DeviceClasses" term_id="deviceclass" >}}, {{< glossary_tooltip text="ResourceSlices" term_id="resourceslice" >}}, {{< glossary_tooltip text="ResourceClaims" term_id="resourceclaim" >}}, as well as new fields in the Pod spec itself.
<!-- objectives -->
{{% heading "objectives" %}}
- Deploy an example DRA driver
- Deploy a Pod requesting a hardware claim using DRA APIs
- Delete a Pod that has a claim
<!-- prerequisites -->
{{% heading "prerequisites" %}}
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.
{{< include "task-tutorial-prereqs.md" >}}
This tutorial has been tested with Linux nodes, though it may also work with other types of nodes.
{{< version-check >}}
If your cluster is not currently running Kubernetes {{< skew currentVersion >}} then please check the documentation for the version of Kubernetes that you plan to use.
<!-- lessoncontent -->
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.
- Get a list of {{< glossary_tooltip text="DeviceClasses" term_id="deviceclass" >}}:
kubectl get deviceclassesThe output is similar to this:
No resources found- Get a list of {{< glossary_tooltip text="ResourceSlices" term_id="resourceslice" >}}:
kubectl get resourceslicesThe output is similar to this:
No resources found- Get a list of {{< glossary_tooltip text="ResourceClaims" term_id="resourceclaim" >}} and
{{< glossary_tooltip text="ResourceClaimTemplates" term_id="resourceclaimtemplate" >}}.
kubectl get resourceclaims -A
kubectl get resourceclaimtemplates -AThe output is similar to this:
No resources found
No resources foundAt 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 {{< glossary_tooltip term_id="daemonset" >}} to all or a selection of the nodes (using {{< glossary_tooltip text="selectors" term_id="selector" >}} 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:
- Create the namespace:
kubectl create namespace dra-tutorialIn 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.
- 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:v0.2.0Deploy the DRA driver components
For this tutorial, you will install the critical example resource driver components individually with kubectl.
- Create the DeviceClass representing the device types this DRA driver
supports:
{{% code_sample language="yaml" file="dra/driver-install/deviceclass.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/deviceclass.yaml- 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:
- Create the Service Account:
{{% code_sample language="yaml" file="dra/driver-install/serviceaccount.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/serviceaccount.yaml- Create the ClusterRole:
{{% code_sample language="yaml" file="dra/driver-install/clusterrole.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/clusterrole.yaml- Create the ClusterRoleBinding:
{{% code_sample language="yaml" file="dra/driver-install/clusterrolebinding.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/clusterrolebinding.yaml- Create a {{< glossary_tooltip term_id="priority-class" >}} 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.
{{% code_sample language="yaml" file="dra/driver-install/priorityclass.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/priorityclass.yaml- 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.
{{% code_sample language="yaml" file="dra/driver-install/daemonset.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/daemonset.yamlThe 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}
- 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-tutorialThe 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- 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 resourceslicesThe 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 19sAt 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 {{< glossary_tooltip term_id="cel" >}} 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 {{< glossary_tooltip term_id="cel" >}} 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.
{{% code_sample language="yaml" file="dra/driver-install/example/resourceclaim.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/example/resourceclaim.yamlCreate 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.
{{% code_sample language="yaml" file="dra/driver-install/example/pod.yaml" %}}
kubectl apply --server-side -f https://k8s.io/examples/dra/driver-install/example/pod.yamlConfirm the pod has deployed:
kubectl get pod pod0 -n dra-tutorialThe output is similar to this:
NAME READY STATUS RESTARTS AGE
pod0 1/1 Running 0 9sExplore 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.
- 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"- Check the state of the ResourceClaim object:
kubectl get resourceclaims -n dra-tutorialThe output is similar to this:
NAME STATE AGE
some-gpu allocated,reserved 34sIn this output, the STATE column shows that the ResourceClaim is allocated and reserved.
- Check the details of the
some-gpuResourceClaim. Thestatusstanza 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 yamlThe output is similar to this:
{{< highlight yaml "linenos=inline, hl_lines=27-30 38-41, style=emacs" >}} 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'))
توجه: = 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 {{< /highlight >}}
- 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-tutorialThe 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-tutorialThe output is similar to this:
pod "pod0" deletedObserve 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.
- Check the state of the
some-gpuResourceClaim:
kubectl get resourceclaims -n dra-tutorialThe output is similar to this:
NAME STATE AGE
some-gpu pending 76s- 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-tutorialThe output is similar to this:
I0820 18:22:15.629376 1 driver.go:138] UnprepareResourceClaims is called: number of claims: 1You 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.
{{% heading "cleanup" %}}
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{{% heading "whatsnext" %}}
- Learn more about DRA
- Allocate Devices to Workloads with DRA