Skip to content

chore(deps): update CUDA LTS image for NVIDIA toolkit 1.20.1 - #9834

Merged
Cameron Meissner (cameronmeissner) merged 2 commits into
mainfrom
bump/cuda-lts-toolkit-1.20.1
Oct 20, 2026
Merged

Cameron Meissner (cameronmeissner) merged 2 commits into
mainfrom
bump/cuda-lts-toolkit-1.20.1

Conversation

@sulixu

@sulixu sulixu commented Oct 20, 2026

Copy link
Copy Markdown
Contributor

What this PR does / why we need it:

Update only the aks-gpu-cuda-lts image pin in parts/common/components.json:

580.159.04-20260918140533 -> 580.159.04-20261020182457

The published image contains NVIDIA Container Toolkit 1.20.1-1, including the change from Azure/aks-gpu#177. The NVIDIA driver version stays at 580.159.04.

This pin controls the CUDA LTS image cached during Ubuntu VHD builds and selected for managed GPU installation during node provisioning. It applies to CUDA LTS nodes, including Standard_NC8as_T4_v3; it is not limited to that SKU.

This is a manual update because the latest AgentBaker Renovate job timed out. No Renovate configuration, GRID image pin, legacy CUDA image pin, Azure Linux package pin, or installation script is changed.

The CUDA LTS publication retry succeeded, and the image is available in MCR:
https://github.com/Azure/aks-gpu/actions/runs/37664513588/job/118806612266

Which issue(s) this PR fixes:

Consumes Azure/aks-gpu#177. No issue is automatically closed.

Validation:

  • git diff --check: passed.
  • JSON validation and comparison against the base revision: only the CUDA LTS image tag changed.
  • cue vet -c ./schemas/components.cue ./parts/common/components.json: passed.
  • go test -mod=readonly ./pkg/agent/datamodel ./pkg/agent -run '^(TestLoadConfig|TestGPUGridDriverBranches|TestGPUImageRepo|TestGPUDriverSelection)$' -count=1: passed.
  • A standalone check using AgentBaker's GPU-selection functions confirmed that Standard_NC8as_T4_v3 selects mcr.microsoft.com/aks/aks-gpu-cuda-lts:580.159.04-20261020182457.
  • Confirmed the MCR image index contains both linux/amd64 and linux/arm64 manifests.

The Go checks used the existing Go 1.26.6 toolchain. The configured dependency proxy returned HTTP 401, so the successful retry used a command-local public Go proxy setting; no saved Go configuration was changed.

No VHD build or live GPU test was run locally.

Copilot AI balanced review requested due to automatic review settings October 20, 2026 19:42
@github-actions github-actions Bot added the components This pull request updates cached components on Linux or Windows VHDs label Oct 20, 2026
@github-actions

github-actions Bot commented Oct 20, 2026 •

Copy link
Copy Markdown
Contributor

Windows Unit Test Results

  3 files   18 suites   1m 14s ⏱️
563 tests 563 ✅ 0 💤 0 ❌
585 runs  585 ✅ 0 💤 0 ❌

Results for commit 774e587.

♻️ This comment has been updated with latest results.

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

🔵 Needs a closer look

The GPU toolkit update affects runtime behavior without reported VHD-build or live GPU validation, so human review is needed.

No open issues 🎉

What changed in this PR

Updates AgentBaker’s CUDA LTS image used for Ubuntu VHD caching and managed GPU provisioning.

Changes:

  • Advances the image tag from 580.159.04-20260918140533 to 580.159.04-20261020182457.
  • Consumes NVIDIA Container Toolkit 1.20.1-1, keeping driver version 580.159.04 unchanged.
  • Leaves other image pins and installation scripts unchanged.
File Description
parts/​common/​components.json Updates the CUDA LTS image pin.

Review effort: Balanced

Copilot AI balanced review requested due to automatic review settings October 20, 2026 20:22
@github-actions

Copy link
Copy Markdown
Contributor

Code Coverage

Package Line Rate Health
test 0% ❌
windows 27% ❌
staging_cse_windows Package 1 65% ➖
debug 0% ❌
provisioningscripts 2% ❌
parts_windows Package 1 73% ➖
Summary 38% (1652 / 6834) ❌

@github-actions

Copy link
Copy Markdown
Contributor

Code Coverage

Package Line Rate Health
github.com/Azure/agentbaker/aks-node-controller 72% ➖
github.com/Azure/agentbaker/aks-node-controller/common 100% ✔
github.com/Azure/agentbaker/aks-node-controller/helpers 68% ➖
github.com/Azure/agentbaker/aks-node-controller/parser 88% ✔
github.com/Azure/agentbaker/aks-node-controller/pkg/gpu 100% ✔
github.com/Azure/agentbaker/aks-node-controller/pkg/nodeconfigutils 88% ✔
github.com/Azure/agentbaker/aks-node-controller/utils 0% ❌
github.com/Azure/agentbaker/apiserver 24% ❌
github.com/Azure/agentbaker/cmd 0% ❌
github.com/Azure/agentbaker/cmd/starter 0% ❌
github.com/Azure/agentbaker/fuzz/api 0% ❌
github.com/Azure/agentbaker/hotfix/render-nodecustomdata 0% ❌
github.com/Azure/agentbaker/pkg/agent 77% ➖
github.com/Azure/agentbaker/pkg/agent/datamodel 78% ➖
github.com/Azure/agentbaker/pkg/agent/toggles 0% ❌
github.com/Azure/agentbaker/pkg/vhdbuilder/datamodel 87% ✔
Summary 74% (5660 / 7595) ➖

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

🔵 Needs a closer look

The toolkit update affects GPU runtime behavior without reported VHD-build or live GPU validation.

No open issues 🎉

Review effort: Balanced

@github-actions

Copy link
Copy Markdown
Contributor

Code Coverage

Package Line Rate Health
shellspec spec 28% ❌
Summary 28% (4507 / 15936) ❌

@sulixu
sulixu enabled auto-merge (squash) October 20, 2026 21:08
@cameronmeissner
Cameron Meissner (cameronmeissner) merged commit 74c56df into main Oct 20, 2026
47 of 51 checks passed
@cameronmeissner
Cameron Meissner (cameronmeissner) deleted the bump/cuda-lts-toolkit-1.20.1 branch October 20, 2026 22:11
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

components This pull request updates cached components on Linux or Windows VHDs

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants