fix(compute-plane): bump nvca to 3.2.6 - #980
Conversation
The stack pinned nvca-operator and NVCA at 3.2.0 while the v3.2 patch line is at 3.2.6. 3.2.6 carries the Helm LLM transport trust backport, which self-managed clusters need when addons.llm.pki is enabled: without it, Helm-based LLM workloads cannot verify the PKI-issued stargate QUIC certificate against the cluster trust bundle. Golden testdata regenerated with make generate-golden. Besides the version bumps, regeneration normalizes one line in agent-config-merge-cm.yaml (the merged config block scalar). That drift predates this change: a fresh render of the unmodified stack produces the same line, so the committed golden was stale. Verified on a self-managed k3d cluster from main with addons.llm.pki enabled and nvca 3.2.6: a container LLM function deploys to ACTIVE and serves invocations over the TLS QUIC path. make test-local passes. Signed-off-by: along <along@nvidia.com>
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: ⛔ Files ignored due to path filters (4)
📒 Files selected for processing (1)
Included review availability: Your plan provides up to 12 included reviews per hour; 10 remain after this review. 📝 WalkthroughWalkthroughThe base environment updates the NVCA operator image tag and self-managed NVCA version from ChangesNVCA version update
Estimated code review effort: 1 (Trivial) | ~2 minutes Merge Risk: ⚪ Minimal · up to This change updates the compute-plane NVCA versions and matching generated outputs without introducing an actionable merge-blocking risk; it is ready for normal checks and review. Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
TL;DR
Bump the compute-plane stack's nvca-operator and NVCA pins from 3.2.0 to 3.2.6 and regenerate the golden testdata. 3.2.6 carries the Helm LLM transport trust backport that self-managed clusters need when addons.llm.pki is enabled.
Additional Details (optional for docs, build, test, refactor, ci, chore, style, and revert PRs)
Regenerating the goldens also normalizes one line in agent-config-merge-cm.yaml (the merged config block scalar header). That drift predates this change: a fresh render of the unmodified stack produces the same line, so the committed golden was stale and this PR brings it back in sync with what make generate-golden produces today.
For the Reviewer
The functional change is two values in deploy/stacks/nvcf-compute-plane/environments/base.yaml. Everything else is regenerated golden testdata from make generate-golden.
For QA (optional for docs, build, test, refactor, ci, chore, style, and revert PRs)
Issues
Relates to #52
Checklist