fix(storage): [backport v3.2] ensure model-cache init namespace has unbound DNS label in NVMesh path - #1304
Merged
apartha-nv merged 7 commits intoAug 31, 2026
Conversation
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. 🗂️ Base branches to auto review (1)
Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Comment |
apartha-nv
force-pushed
the
backport/1303-modelcache-nvmesh-unbound-dns-v3.2
branch
from
August 28, 2026 09:34
0798ab0 to
e1116ad
Compare
sbaum1994
approved these changes
Aug 31, 2026
…space at startup The namespace is created once at NVCA startup in backendk8scache.go with AlreadyExists-is-success and no patch, so pre-existing namespaces on upgraded clusters never received WorkloadInstanceTypeLabel regardless of backend -- the samba-path fix (#1116) only applied during a model-attached deploy and only for the samba backend. Add ensureModelCacheNamespaceLabel immediately after the Create call in backendk8scache.go. It runs on every NVCA restart via JSON patch 'add' (idempotent: inserts when absent, updates when present), so upgraded clusters receive the label immediately without needing a model-attached helm deploy to trigger reconciliation. The samba path's ensureNamespaceLabels remains as belt-and-suspenders for the case where samba creates the namespace itself. Relates to NO-REF
… label already present
…l; add envtest JSON patch 'add' requires the parent path /metadata/labels to exist; a namespace with nil labels would cause startup to fail with a patch error. Switch to strategic merge patch which creates the labels map when absent and merges into it when present. Add envtest covering: - nil labels (the case JSON patch 'add' would have rejected) - pre-existing labels (merge preserves unrelated keys) - label already correct (idempotent, no error) Update mock tests to match the new patch type and payload.
…test deps Required by TestEnsureModelCacheNamespaceLabel_Envtest which uses nvcaenvtest.SetupEnvtest and kubernetes.NewForConfig.
Tests using nvcaenvtest.SetupEnvtest must include the CRD manifests as data so Bazel materializes them in the runfiles tree at the path runtime.Caller(0) resolves to in envtest.go.
…NVCA CRDs needed nvcaenvtest.SetupEnvtest loads NVCA CRDs from a path resolved via runtime.Caller(0), which breaks in Bazel sandboxes because the source tree is not present at that path. The namespace label test only needs core Kubernetes resources (Namespace), so switch to a plain envtest.Environment without CRD loading. Also revert the now-unused internal/envtest dep and CRD data entry from BUILD.bazel.
apartha-nv
force-pushed
the
backport/1303-modelcache-nvmesh-unbound-dns-v3.2
branch
from
August 31, 2026 19:43
3623c4d to
35f7ac8
Compare
apartha-nv
merged commit Aug 31, 2026
a967088
into
release-src/compute-plane-services/nvca/v3.2
12 checks passed
apartha-nv
deleted the
backport/1303-modelcache-nvmesh-unbound-dns-v3.2
branch
August 31, 2026 20:07
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Customer Summary
Backport of #1303 to v3.2.
Fixes model-cache initialization for helm-chart functions on NVMesh clusters (e.g. forge/DGXC). Confirmed still reproducing on 3.2.15 — the samba-path fix from #1218 did not cover NVMesh clusters.
TL;DR
Cherry-pick of #1303 onto
release-src/compute-plane-services/nvca/v3.2. AddsensureCreated + ensureNamespaceLabelsatdoInitModelCacheNVMeshentry so the namespace getsWorkloadInstanceTypeLabelon NVMesh clusters.Tickets
Backport of #1303. Follow-up to #1218. Relates to NO-REF.