Summary

Disable the Dynamo MPI ssh-keygen Helm hook only for the h100-kind-inference-dynamo recipe.

Motivation / Context

The Kind GPU inference CI workload does not exercise Dynamo's MPI launch path, but the upstream chart still runs a pre-install/pre-upgrade ssh-keygen hook that pulls bitnamisecure/git:latest. In run 24889208608, that hook repeatedly timed out and consumed most of the GPU job budget before the snapshot step was canceled.

Fixes: N/A
Related: https://github.com/NVIDIA/aicr/actions/runs/24889208608/job/72876842190?pr=666

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update
• Refactoring (no functional changes)
• Build/CI/tooling

Component(s) Affected

• CLI (cmd/aicr, pkg/cli)
• API server (cmd/aicrd, pkg/api, pkg/server)
• Recipe engine / data (pkg/recipe)
• Bundlers (pkg/bundler, pkg/component/*)
• Collectors / snapshotter (pkg/collector, pkg/snapshotter)
• Validator (pkg/validator)
• Core libraries (pkg/errors, pkg/k8s)
• Docs/examples (docs/, examples/)
• Other: ____________

Implementation Notes

Adds an inline override under the Kind Dynamo leaf recipe:

dynamo-operator:
  dynamo:
    mpiRun:
      sshKeygen:
        enabled: false

This is intentionally scoped to h100-kind-inference-dynamo; cloud Dynamo overlays continue to use the chart default.

Testing

go test ./pkg/recipe/...
make lint-yaml
aicr query --service kind --accelerator h100 --intent inference --platform dynamo --data recipes --selector components.dynamo-platform.values.dynamo-operator.dynamo.mpiRun.sshKeygen.enabled
unset GITLAB_TOKEN && make qualify

Results:

• go test ./pkg/recipe/... passed
• make lint-yaml passed
• Query returned false
• make qualify passed, including race tests, coverage, lint, e2e, scan, and license checks

Risk Assessment

• Low — Isolated change, well-tested, easy to revert
• Medium — Touches multiple components or has broader impact
• High — Breaking change, affects critical paths, or complex rollout

Rollout notes: N/A

Checklist

• Tests pass locally (make test with -race)
• Linter passes (make lint)
• I did not skip/disable tests to make CI green
• I added/updated tests for new functionality
• I updated docs if user-facing behavior changed
• Changes follow existing patterns in the codebase
• Commits are cryptographically signed (git commit -S) — GPG signing info