Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

✨ vGPU implementation #2272

Closed
wants to merge 2 commits into from
Closed

✨ vGPU implementation #2272

wants to merge 2 commits into from

Conversation

puneetkatyal
Copy link
Contributor

What this PR does / why we need it:
Support adding vGPUs to VMs

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
Fixes #1972

@k8s-ci-robot k8s-ci-robot added the cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. label Aug 22, 2023
@k8s-ci-robot
Copy link
Contributor

Hi @puneetkatyal. Thanks for your PR.

I'm waiting for a kubernetes-sigs member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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/test-infra repository.

@k8s-ci-robot k8s-ci-robot added the needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. label Aug 22, 2023
@k8s-ci-robot
Copy link
Contributor

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign neolit123 for approval. For more information see the Kubernetes Code Review Process.

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

Needs approval from an approver in each of these files:

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

@k8s-ci-robot k8s-ci-robot added the size/L Denotes a PR that changes 100-499 lines, ignoring generated files. label Aug 22, 2023
@sbueringer
Copy link
Member

/ok-to-test

@k8s-ci-robot k8s-ci-robot added ok-to-test Indicates a non-member PR verified by an org member that is safe to test. and removed needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. labels Aug 23, 2023
@puneetkatyal
Copy link
Contributor Author

/retest

@puneetkatyal @geetikabatra
vGPU implementation
d3615dd 
- Builds on the changes in
  kubernetes-sigs#1579

Co-authored-by: Geetika Batra <[email protected]>
Signed-off-by: Puneet Katyal <[email protected]>
chrischdi
chrischdi reviewed Aug 28, 2023
View reviewed changes
Copy link
Member

@chrischdi chrischdi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There are some general questions for me:

  • I think we should prefer only keeping reconcilePCIDevices as the way for the normal PCI Passthrough. Especially because this also handles some condition handling.
  • Question: instead of adding vgpu devices directly at clone: should this maybe also be done inside reconcilePCIDevices? (because they are just a bit more special PCI devices?)
  • API conversion

Note: I'm not able to verify that all of this works because I currently don't have a environment including vgpu's available.

apis/v1alpha3/zz_generated.conversion.go
@@ -1685,6 +1685,7 @@ func autoConvert_v1beta1_VirtualMachineCloneSpec_To_v1alpha3_VirtualMachineClone
out.CustomVMXKeys = *(*map[string]string)(unsafe.Pointer(&in.CustomVMXKeys))
// WARNING: in.TagIDs requires manual conversion: does not exist in peer-type
// WARNING: in.PciDevices requires manual conversion: does not exist in peer-type
// WARNING: in.VGPUDevices requires manual conversion: does not exist in peer-type
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we have to implement conversion for this.

docs/gpu-vgpu.md
template: '${VSPHERE_TEMPLATE}'
thumbprint: '${VSPHERE_TLS_THUMBPRINT}'
vgpuDevices:
- profileName: "grid_v100d-4c" <============ value from above
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To make it a valid yaml

Suggested change
- profileName: "grid_v100d-4c" <============ value from above 
    - profileName: "grid_v100d-4c" # value from above!

docs/gpu-vgpu.md
Comment on lines +59 to +60
/Applications/Xcode.app/Contents/Developer/usr/bin/make generate-flavors FLAVOR_DIR=/Users/pkatyal/.cluster-api/overrides/infrastructure-vsphere/v0.0.0
go run ./packaging/flavorgen --output-dir /Users/pkatyal/.cluster-api/overrides/infrastructure-vsphere/v0.0.0
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
/Applications/Xcode.app/Contents/Developer/usr/bin/make generate-flavors FLAVOR_DIR=/Users/pkatyal/.cluster-api/overrides/infrastructure-vsphere/v0.0.0
go run ./packaging/flavorgen --output-dir /Users/pkatyal/.cluster-api/overrides/infrastructure-vsphere/v0.0.0

Let's omit the stdout output

docs/gpu-vgpu.md
template: '${VSPHERE_TEMPLATE}'
thumbprint: '${VSPHERE_TLS_THUMBPRINT}'
vgpuDevices:
- profileName: "grid_v100d-4c" <============ value from above
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would it be worth having this as envsubst parameter in some way?

WDYT: would it be worth having a separate flavor for vgpu?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The same NVIDIA GPU supports multiple vGPU profiles, and the matrix expands when you add more GPU varieties to the mix. For example, in my testing, I use different profiles for different worker nodes for the same workload cluster. I don't think it's useful to have this as an envsubst parameter.

pkg/services/govmomi/vcenter/clone.go
@@ -453,3 +461,55 @@ func getNetworkSpecs(ctx *context.VMContext, devices object.VirtualDeviceList) (

return deviceSpecs, nil
}

func createPCIPassThroughDevice(deviceKey int32, backingInfo types.BaseVirtualDeviceBackingInfo) types.BaseVirtualDevice {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's move this func to the very bottom of the file, because its a helper for the ones below.

pkg/services/govmomi/vcenter/clone.go Outdated Show resolved Hide resolved
pkg/services/govmomi/vcenter/clone.go
Device: dynamicDirectPathDevice,
Operation: types.VirtualDeviceConfigSpecOperationAdd,
})
ctx.Logger.V(4).Info("created vGPU device", "vgpu-profile", vGPUDevice.ProfileName)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should we have a smilar message for the gpu case?

pkg/services/govmomi/vcenter/clone.go
return device
}

func getGpuSpecs(ctx *context.VMContext) []types.BaseVirtualDeviceConfigSpec {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could we try to not use the *context.VMContext parameter here and use more specific parameters instead?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also: this function seems to be a duplicate of https://github.com/chrischdi/cluster-api-provider-vsphere/blob/234e7751f7bb09bb438eed6b4feabbc274cae842/pkg/services/govmomi/pci/device.go#L59 ?

pkg/services/govmomi/vcenter/clone.go
@@ -151,8 +151,16 @@ func Clone(ctx *context.VMContext, bootstrapData []byte, format bootstrapv1.Form

deviceSpecs = append(deviceSpecs, networkSpecs...)

if err != nil {
return errors.Wrapf(err, "error getting network specs for %q", ctx)
if len(ctx.VSphereVM.Spec.VirtualMachineCloneSpec.PciDevices) != 0 {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This adds a second codepath which adds PciDevices to a VM. The already existing code is at https://github.com/chrischdi/cluster-api-provider-vsphere/blob/234e7751f7bb09bb438eed6b4feabbc274cae842/pkg/services/govmomi/service.go#L478 .

I'd prefer only having one place to do this so either this here should get removed or the other place. And we have to take care to not break existing use cases.

@k8s-ci-robot k8s-ci-robot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Oct 13, 2023
@k8s-ci-robot
Copy link
Contributor

PR needs rebase.

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/test-infra repository.

@k8s-triage-robot
Copy link

The Kubernetes project currently lacks enough contributors to adequately respond to all PRs.

This bot triages PRs according to the following rules:

  • After 90d of inactivity, lifecycle/stale is applied
  • After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
  • After 30d of inactivity since lifecycle/rotten was applied, the PR is closed

You can:

  • Mark this PR as fresh with /remove-lifecycle stale
  • Close this PR with /close
  • Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

@k8s-ci-robot k8s-ci-robot added the lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. label Jan 22, 2024
@sbueringer
Copy link
Member

/close

Given no changes in the last ~ 5 months. Feel free to reopen if you want to continue working on it.

@k8s-ci-robot
Copy link
Contributor

@sbueringer: Closed this PR.

In response to this:

/close

Given no changes in the last ~ 5 months. Feel free to reopen if you want to continue working on it.

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/test-infra repository.

@birksl birksl mentioned this pull request May 24, 2024
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@chrischdi chrischdi chrischdi left review comments

@gab-satchi gab-satchi Awaiting requested review from gab-satchi

@MaxRink MaxRink Awaiting requested review from MaxRink

Assignees
No one assigned
Labels
cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. ok-to-test Indicates a non-member PR verified by an org member that is safe to test. size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

vGPU support
5 participants
@puneetkatyal @k8s-ci-robot @sbueringer @k8s-triage-robot @chrischdi