Enable DRA device metadata (KEP-5304)

[carlory]

Summary

Add support for KEP-5304 in-container device metadata by passing device attributes from the allocatable device pool through to prepared devices.

• New --enable-device-metadata CLI flag and ENABLE_DEVICE_METADATA env var
• New kubeletPlugin.enableDeviceMetadata Helm chart value
• Device attributes from the driver's allocatable pool are propagated as metadata on prepared devices
• Bumps Go to 1.26.0 and all k8s.io dependencies to v0.36.0-rc.1

/cc @alaypatel07
I add you as co-author. If you want to submit a PR, feel free to close this PR.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: carlory
Once this PR has been reviewed and has the lgtm label, please assign bart0sh for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: alaypatel07 / name: Alay Patel (1c15898)
• ✅ login: carlory / name: carlory (1c15898, b032b4e)

[nojnhuh]

I have a draft of the changes to bump to 1.36 here but haven't opened a PR yet: main...nojnhuh:dra-example-driver:k8s-1.36

I'm ok if we want to do that in this PR, but can we at least make those changes a separate commit here? There are also some other places that need to be updated here that you can check against my branch.

[carlory]

v1.36 will be released in two days, once it is released, I will update this PR.

[nojnhuh]

Can we update the e2e tests to enable and exercise this flag?

#153 refactors all of the e2e logic but I think that will merge soon, so you may want to hold off on that until that merges.

[carlory]

v1.36.0 is unavailable.

~ » skopeo list-tags docker://docker.io/kindest/node | jq '.Tags[-1]'                                          kiki@zhanxis-MacBook-Pro
"v1.35.1"

[nojnhuh]

We could set BUILD_KIND_IMAGE based on whether $KIND_IMAGE exists instead of a static true/false.

dra-example-driver/demo/scripts/common.sh

Lines 41 to 44 in a538bc8

	# At present, kind has a new enough node image that we don't need to build our
	# own. This won't always be true and we may need to set the variable below to
	# 'true' from time to time as things change.
	: ${BUILD_KIND_IMAGE:="false"}

I've noticed resource constraint issues with building the kind image from source in CI. Then we could refactor demo/scripts/build-kind-image.sh to allow building the image from release artifacts. We would need a way to pass --type release <version> here instead of a file path to the cloned repo:

dra-example-driver/demo/scripts/build-kind-image.sh

Line 55 in a538bc8

${KIND} build node-image --image "${KIND_IMAGE}" "${KIND_K8S_DIR}"

A generic variable like KIND_BUILD_NODE_IMAGE_ARGS is probably good enough to handle a variety of use cases.

[nojnhuh]

I'm trying this out in #188.

[k8s-ci-robot]

PR needs rebase.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.