Skip to content

Commit e295f90

Browse files
authored
feat(kubernetes): propagate remote-accessible LINSTOR StorageClasses to tenant clusters (#2872)
## What this PR does Tenant Kubernetes clusters previously exposed a single hardcoded `kubevirt` StorageClass mapped to one infrastructure class (`.Values.storageClass`). Users running multiple storage tiers on the management cluster could not reach them from inside tenant clusters (#2094, #1424). This change auto-discovers the management cluster's **remote-accessible LINSTOR** StorageClasses (`provisioner: linstor.csi.linbit.com` with `allowRemoteVolumeAccess != "false"`) via Helm `lookup` and propagates each to tenant clusters under the same name. No Go changes — the kubevirt-csi-driver already supports multiple infra StorageClasses natively. **Behavior** - `packages/apps/kubernetes` `csi.yaml` builds a `storageClasses` map filtered to remote-accessible LINSTOR classes and passes it to the `kubevirt-csi-node` child chart. - `kubevirt-csi-node` renders one tenant StorageClass per entry (`provisioner: csi.kubevirt.io`, `infraStorageClassName: <name>`, `bus: scsi`). - The class named by `.Values.storageClass` (default `replicated`) is marked the tenant default; exactly one default is guaranteed. - The legacy `kubevirt` StorageClass is retained as a non-default alias, so **existing tenant PVCs keep binding with no migration**. - If the `lookup` returns empty (e.g. dry-run), a non-destructive fallback preserves the legacy `replicated` + `kubevirt` mapping — tenant StorageClasses are never wiped on a transient reconcile. **Scope decisions** - **LINSTOR-only**: only StorageClasses cozystack ships and controls are auto-propagated. Node-local classes (e.g. `local`, `allowRemoteVolumeAccess: false`) are excluded because the backing volume cannot follow a live-migrated worker-node VM. Other provisioners (NFS, etc.) are out of scope. - **No enforcement flip**: `infraStorageClassEnforcement` is intentionally left at the current `AllowAll` default. Enabling the allowList would break tenants that have already self-served custom StorageClasses against the management cluster — exactly the users this feature serves. Tightening tenant isolation is left to a separate, discovery-based change. This is a focused, additive alternative to #2095 (kept open), scoped per the LINSTOR-only / no-breakage decisions above. **Upgrade note** If you previously created a tenant StorageClass **named `replicated`** by hand, delete it before upgrading — otherwise the child Helm release hits an ownership conflict adopting the now-managed `replicated` class. ### Tests - Helm unit tests for the fallback map (`kubernetes`) and per-entry rendering + single-default invariant + absence of an enforcement block (`kubevirt-csi-node`; new `Makefile` `test` target so the package is no longer skipped by the runner). - e2e (`run-kubernetes.sh`): asserts `replicated` propagates as the default kubevirt-CSI class, the `kubevirt` alias is retained, and node-local `local` is filtered out. ### Release note ```release-note feat(kubernetes): auto-propagate remote-accessible LINSTOR StorageClasses from the management cluster to tenant Kubernetes clusters under the same name. The class set by `storageClass` (default `replicated`) becomes the tenant default; the legacy `kubevirt` StorageClass is kept as an alias so existing PVCs are unaffected. Node-local classes (e.g. `local`) are not propagated. If you manually created a tenant StorageClass named `replicated`, delete it before upgrading to avoid a Helm ownership conflict. ``` <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **New Features** * Remote-accessible LINSTOR `StorageClass` entries now auto-propagate into tenant clusters using the same name. * Tenant-side CSI/KubeVirt `StorageClass` rendering now supports multiple configurable classes with deterministic default selection. * **Improvements** * Enhanced e2e checks to verify `StorageClass` provisioner, parameters, default annotation, and legacy `kubevirt` alias compatibility. * **Tests** * Added Helm unittest suites covering fallback and explicit `storageClass` cases. * Added kubevirt CSI node rendering tests and a `test` target. * **Documentation** * Clarified `storageClass` descriptions and upgrade caveats across charts and schemas. <!-- end of auto-generated comment: release notes by coderabbit.ai -->
2 parents ee54a6e + da084ef commit e295f90

13 files changed

Lines changed: 376 additions & 13 deletions

File tree

‎api/apps/v1alpha1/kubernetes/types.go‎

Lines changed: 1 addition & 1 deletion
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

‎hack/e2e-apps/kubernetes-latest.bats‎

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -11,3 +11,11 @@
1111
# spinning up a second ~25m Kamaji bringup.
1212
run_kubernetes_test 'keys | sort_by(.) | .[-1]' 'test-latest-version' '59991' 'true'
1313
}
14+
15+
# Version-independent, so it runs once here (not in kubernetes-previous.bats).
16+
# Does not provision a tenant cluster -- it renders the kubernetes chart against
17+
# the live management cluster -- so it is kept separate from the heavy tenant
18+
# test above and runs regardless of that test's outcome.
19+
@test "Tenant default StorageClass is never the kubevirt alias (PR #2872 B1 regression)" {
20+
verify_storageclass_fallback_default
21+
}

‎hack/e2e-apps/run-kubernetes.sh‎

Lines changed: 170 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -383,6 +383,50 @@ EOF
383383
# the 5m pod-Succeeded budget when containerd's CreateContainer stalls.
384384
kubectl wait hr -n tenant-test "kubernetes-${test_name}-csi" --timeout=10m --for=condition=ready
385385

386+
# ----------------------------------------------------------------------
387+
# StorageClass propagation (issue #2094). Remote-accessible LINSTOR infra
388+
# classes propagate to the tenant under the same name; node-local classes
389+
# ("local", allowRemoteVolumeAccess=false) are filtered out; the legacy
390+
# "kubevirt" alias is retained for backward compatibility. The e2e infra
391+
# cluster ships both "replicated" (remote) and "local" (node-local).
392+
# ----------------------------------------------------------------------
393+
echo "Verifying StorageClass propagation to tenant..."
394+
timeout 2m bash -c '
395+
until kubectl --kubeconfig tenantkubeconfig-'"${test_name}"' get sc replicated >/dev/null 2>&1; do
396+
sleep 5
397+
done
398+
'
399+
400+
rep_prov=$(kubectl --kubeconfig "tenantkubeconfig-${test_name}" get sc replicated -o jsonpath='{.provisioner}')
401+
rep_infra=$(kubectl --kubeconfig "tenantkubeconfig-${test_name}" get sc replicated -o jsonpath='{.parameters.infraStorageClassName}')
402+
if [ "$rep_prov" != "csi.kubevirt.io" ] || [ "$rep_infra" != "replicated" ]; then
403+
echo "replicated SC misconfigured: provisioner=$rep_prov infraStorageClassName=$rep_infra" >&2
404+
kubectl --kubeconfig "tenantkubeconfig-${test_name}" get sc >&2
405+
exit 1
406+
fi
407+
408+
# Legacy kubevirt alias must still exist (existing PVCs depend on it).
409+
if ! kubectl --kubeconfig "tenantkubeconfig-${test_name}" get sc kubevirt >/dev/null 2>&1; then
410+
echo "legacy kubevirt StorageClass alias is missing" >&2
411+
exit 1
412+
fi
413+
414+
# Node-local "local" class must NOT be propagated (allowRemoteVolumeAccess=false).
415+
if kubectl --kubeconfig "tenantkubeconfig-${test_name}" get sc local >/dev/null 2>&1; then
416+
echo "node-local StorageClass 'local' should not be propagated to the tenant" >&2
417+
exit 1
418+
fi
419+
420+
# Exactly one default StorageClass, and it must be "replicated".
421+
default_scs=$(kubectl --kubeconfig "tenantkubeconfig-${test_name}" get sc \
422+
-o jsonpath='{range .items[?(@.metadata.annotations.storageclass\.kubernetes\.io/is-default-class=="true")]}{.metadata.name}{"\n"}{end}')
423+
default_count=$(printf '%s' "$default_scs" | grep -c .)
424+
if [ "$default_count" -ne 1 ] || [ "$default_scs" != "replicated" ]; then
425+
echo "expected exactly one default StorageClass 'replicated', got: ${default_scs:-<none>} (count=$default_count)" >&2
426+
exit 1
427+
fi
428+
echo "StorageClass propagation OK (replicated default, kubevirt alias present, local filtered)"
429+
386430
# Clean up NFS test resources from any previous failed attempt
387431
kubectl --kubeconfig "tenantkubeconfig-${test_name}" delete pod nfs-test-pod \
388432
-n tenant-test --ignore-not-found --timeout=60s || true
@@ -621,3 +665,129 @@ EOF
621665
kubectl -n tenant-test delete kuberneteses.apps.cozystack.io "${test_name}" --ignore-not-found --wait=false 2>/dev/null || true
622666

623667
}
668+
669+
# B1 regression coverage (PR #2872 review). The tenant's default StorageClass
670+
# must be chosen among the *propagated* classes and must never be the legacy
671+
# "kubevirt" alias -- even when the management cluster exposes only remote
672+
# LINSTOR classes whose names sort alphabetically after "kubevirt" and none is
673+
# named the configured storageClass (default "replicated"). That is the
674+
# feature's own multi-tier target configuration. A regressed `sortAlpha | first`
675+
# over a candidate set that still contained the inserted "kubevirt" alias would
676+
# pick it, pointing the tenant default at an infra class absent on the
677+
# management cluster -> default PVCs stay Pending with no error surfaced.
678+
#
679+
# helm-unittest cannot reach this branch: with no live cluster Helm `lookup`
680+
# returns empty, so the storageClasses map always collapses to the "replicated"
681+
# fallback (see packages/apps/kubernetes/tests/csi_test.yaml). It is therefore
682+
# exercised here against the live management cluster with a single server-side
683+
# dry-run render (helm v4 executes `lookup` against the API): add two remote
684+
# LINSTOR classes that sort after "kubevirt", remove "replicated" for the one
685+
# render, restore it immediately, then assert on the rendered -csi HelmRelease's
686+
# storageClasses map.
687+
verify_storageclass_fallback_default() {
688+
echo "Verifying tenant default StorageClass selection with no 'replicated' class (PR #2872 B1 regression)..."
689+
690+
# Pre-cleanup: drop probe classes leaked by a previous failed run.
691+
kubectl delete sc nvme ssd --ignore-not-found
692+
693+
# Two remote-accessible LINSTOR classes whose names sort AFTER "kubevirt".
694+
kubectl apply -f - <<'EOF'
695+
apiVersion: storage.k8s.io/v1
696+
kind: StorageClass
697+
metadata:
698+
name: nvme
699+
provisioner: linstor.csi.linbit.com
700+
parameters:
701+
linstor.csi.linbit.com/storagePool: "data"
702+
linstor.csi.linbit.com/allowRemoteVolumeAccess: "true"
703+
volumeBindingMode: Immediate
704+
allowVolumeExpansion: true
705+
---
706+
apiVersion: storage.k8s.io/v1
707+
kind: StorageClass
708+
metadata:
709+
name: ssd
710+
provisioner: linstor.csi.linbit.com
711+
parameters:
712+
linstor.csi.linbit.com/storagePool: "data"
713+
linstor.csi.linbit.com/allowRemoteVolumeAccess: "true"
714+
volumeBindingMode: Immediate
715+
allowVolumeExpansion: true
716+
EOF
717+
718+
# Remove "replicated" only for the duration of the render below, so that
719+
# neither the configured storageClass (default "replicated") nor "replicated"
720+
# is in the propagated set -- forcing the `sortAlpha | first` selection branch.
721+
kubectl delete sc replicated --ignore-not-found
722+
723+
# Server-side dry-run executes Helm `lookup` against the live cluster and
724+
# renders the real storageClasses map. rc is captured separately (no pipe) so
725+
# the management-cluster state is always restored before any assertion exits.
726+
# The release namespace must be a valid tenant identifier (the chart's
727+
# dashboard-resourcemap template enforces this), so render under tenant-test.
728+
local raw rc
729+
raw=$(timeout 120 helm install scprobe packages/apps/kubernetes \
730+
--dry-run=server -n tenant-test \
731+
-f packages/apps/kubernetes/tests/values/common.yaml -o json 2>/tmp/sc-fallback-render.err)
732+
rc=$?
733+
734+
# Restore management-cluster StorageClasses (inline, unconditional). This MUST
735+
# run before any assertion `exit 1` below, so no EXIT/RETURN trap is used
736+
# (per docs/agents/e2e-testing.md). The "replicated" manifest mirrors
737+
# hack/e2e-post-install-prep.sh.
738+
kubectl apply -f - <<'EOF'
739+
apiVersion: storage.k8s.io/v1
740+
kind: StorageClass
741+
metadata:
742+
name: replicated
743+
provisioner: linstor.csi.linbit.com
744+
parameters:
745+
linstor.csi.linbit.com/storagePool: "data"
746+
linstor.csi.linbit.com/autoPlace: "3"
747+
linstor.csi.linbit.com/layerList: "drbd storage"
748+
linstor.csi.linbit.com/allowRemoteVolumeAccess: "true"
749+
property.linstor.csi.linbit.com/DrbdOptions/auto-quorum: suspend-io
750+
property.linstor.csi.linbit.com/DrbdOptions/Resource/on-no-data-accessible: suspend-io
751+
property.linstor.csi.linbit.com/DrbdOptions/Resource/on-suspended-primary-outdated: force-secondary
752+
property.linstor.csi.linbit.com/DrbdOptions/Net/rr-conflict: retry-connect
753+
volumeBindingMode: Immediate
754+
allowVolumeExpansion: true
755+
EOF
756+
kubectl delete sc nvme ssd --ignore-not-found
757+
758+
if [ "$rc" -ne 0 ] || [ -z "$raw" ]; then
759+
echo "server-side dry-run render of the kubernetes chart failed (rc=$rc)" >&2
760+
cat /tmp/sc-fallback-render.err >&2 || true
761+
exit 1
762+
fi
763+
764+
# Isolate the rendered -csi HelmRelease's storageClasses map.
765+
local sc
766+
sc=$(printf '%s' "$raw" | yq -p=json '.manifest' \
767+
| yq 'select(.kind == "HelmRelease" and .metadata.name == "scprobe-csi") | .spec.values.storageClasses')
768+
if [ -z "$sc" ] || [ "$sc" = "null" ]; then
769+
echo "rendered scprobe-csi HelmRelease carries no storageClasses map" >&2
770+
printf '%s' "$raw" | yq -p=json '.manifest' >&2
771+
exit 1
772+
fi
773+
774+
local default_count default_key kubevirt_present kubevirt_default
775+
default_count=$(printf '%s' "$sc" | yq '[to_entries | .[] | select(.value.default == true)] | length')
776+
default_key=$(printf '%s' "$sc" | yq 'to_entries | map(select(.value.default == true)) | .[0].key')
777+
kubevirt_present=$(printf '%s' "$sc" | yq 'has("kubevirt")')
778+
kubevirt_default=$(printf '%s' "$sc" | yq '.kubevirt.default')
779+
780+
# 1. Exactly one default. 2. The default is a propagated class (nvme/ssd),
781+
# never the kubevirt alias. 3. The kubevirt alias still exists, non-default.
782+
if [ "$default_count" != "1" ] \
783+
|| { [ "$default_key" != "nvme" ] && [ "$default_key" != "ssd" ]; } \
784+
|| [ "$kubevirt_present" != "true" ] \
785+
|| [ "$kubevirt_default" != "false" ]; then
786+
echo "tenant default StorageClass selection regressed (PR #2872 B1):" >&2
787+
echo " default_count=$default_count default_key=$default_key kubevirt_present=$kubevirt_present kubevirt_default=$kubevirt_default" >&2
788+
echo " expected exactly one default among {nvme,ssd}; kubevirt present and non-default" >&2
789+
printf 'rendered storageClasses:\n%s\n' "$sc" >&2
790+
exit 1
791+
fi
792+
echo "StorageClass fallback-default OK (default='$default_key' among propagated classes; kubevirt alias non-default)"
793+
}

‎packages/apps/kubernetes/README.md‎

Lines changed: 5 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -93,15 +93,17 @@ See the reference for components utilized in this service:
9393
9494
- **GitOps-managed tenant `Kubernetes` CRs must bump `spec.version` in Git** (Phase 1 Talos rollover): platform migration 46 patches live `kuberneteses.apps.cozystack.io` objects from `v1.30` to `v1.31` (v1.30 left the Talos↔Kubernetes support matrix). When the tenant `Kubernetes` CR is reconciled from Git (Flux/Argo), the next source reconcile re-applies `version: v1.30` over migration 46's patch, which then trips the chart's `_versions.tpl` guard and the HelmRelease fails. Update `spec.version` to `v1.31` (or newer) in your Git source before/with the platform upgrade. Tenants managed via the API or dashboard need no action — migration 46 handles them.
9595
96+
- **Remote-accessible LINSTOR StorageClasses are auto-propagated to tenants** (v1.5): infra-cluster LINSTOR StorageClasses whose `linstor.csi.linbit.com/allowRemoteVolumeAccess` is not `"false"` are created inside each tenant under the same name (provisioned by `csi.kubevirt.io`). **Upgrade caveat:** delete any *manually created* tenant StorageClass whose name collides with a propagated infra class (e.g. a hand-made `replicated`) before upgrading, or the propagated class cannot be created and the tenant CSI release will not converge. Infra classes that must stay node-local need `allowRemoteVolumeAccess: "false"` set explicitly — an absent annotation is treated as remote-accessible. Propagation is evaluated at HelmRelease render time, so classes added or removed on the infra cluster after a tenant exists propagate only on that tenant's next reconcile.
97+
9698
> The top-level `storageClass` field is annotated as immutable in the chart schema — see [`docs/storage-immutability.md`](../../../docs/storage-immutability.md) for the contract and which consumers enforce it. The per-node-group `nodeGroups[name].storageClass` field is intentionally **not** annotated immutable: it is optional and undefaulted, so a strict `self == oldSelf` rule would block any future attempt to set it on an existing node group.
9799

98100
## Parameters
99101

100102
### Common Parameters
101103

102-
| Name | Description | Type | Value |
103-
| -------------- | ------------------------------------ | -------- | ------------ |
104-
| `storageClass` | StorageClass used to store the data. | `string` | `replicated` |
104+
| Name | Description | Type | Value |
105+
| -------------- | ---------------------------------------------------------------------------------------------------------------------------------------- | -------- | ------------ |
106+
| `storageClass` | Default StorageClass inside the tenant cluster. Remote-accessible LINSTOR classes are auto-propagated to the tenant under the same name. | `string` | `replicated` |
105107

106108

107109
### Application-specific Parameters

‎packages/apps/kubernetes/templates/helmreleases/csi.yaml‎

Lines changed: 28 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -28,10 +28,35 @@ spec:
2828
upgrade:
2929
remediation:
3030
retries: -1
31-
{{- with .Values.storageClass }}
3231
values:
33-
storageClass: "{{ . }}"
34-
{{- end }}
32+
storageClasses:
33+
{{- /* Propagate remote-accessible LINSTOR infra StorageClasses to the tenant under the same name. */}}
34+
{{- $defaultName := .Values.storageClass | default "replicated" }}
35+
{{- $out := dict }}
36+
{{- range $sc := (lookup "storage.k8s.io/v1" "StorageClass" "" "").items }}
37+
{{- $params := $sc.parameters | default dict }}
38+
{{- if and (eq $sc.provisioner "linstor.csi.linbit.com") (ne (index $params "linstor.csi.linbit.com/allowRemoteVolumeAccess" | default "true") "false") }}
39+
{{- $_ := set $out $sc.metadata.name (dict "infraStorageClass" $sc.metadata.name "default" false) }}
40+
{{- end }}
41+
{{- end }}
42+
{{- /* Non-destructive fallback for dry-run/empty lookup: preserve the legacy mapping, never collapse to nothing. */}}
43+
{{- if not $out }}
44+
{{- $_ := set $out "replicated" (dict "infraStorageClass" "replicated" "default" false) }}
45+
{{- end }}
46+
{{- /* Mark exactly one default among the propagated classes. Chosen BEFORE the kubevirt alias is
47+
added below, so the legacy alias (which points at $defaultName, possibly absent on the infra
48+
cluster) can never be auto-selected as the tenant default: the configured storageClass if
49+
propagated, else replicated, else first by name. */}}
50+
{{- $chosen := "" }}
51+
{{- if hasKey $out $defaultName }}{{ $chosen = $defaultName }}
52+
{{- else if hasKey $out "replicated" }}{{ $chosen = "replicated" }}
53+
{{- else }}{{ $chosen = (keys $out | sortAlpha | first) }}{{ end }}
54+
{{- /* Legacy alias so existing tenant PVCs bound to the "kubevirt" StorageClass keep working. */}}
55+
{{- if not (hasKey $out "kubevirt") }}
56+
{{- $_ := set $out "kubevirt" (dict "infraStorageClass" $defaultName "default" false) }}
57+
{{- end }}
58+
{{- $_ := set (index $out $chosen) "default" true }}
59+
{{- $out | toYaml | nindent 6 }}
3560
dependsOn:
3661
- name: {{ .Release.Name }}-vsnap-crd
3762
namespace: {{ .Release.Namespace }}
Lines changed: 60 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,60 @@
1+
suite: csi.yaml StorageClass propagation
2+
3+
templates:
4+
- templates/helmreleases/csi.yaml
5+
6+
values:
7+
- values/common.yaml
8+
9+
# NOTE: helm unittest has no live cluster, so `lookup` returns empty and these
10+
# tests exercise the non-destructive fallback path of the storageClasses map.
11+
# The LINSTOR provisioner / allowRemoteVolumeAccess filtering is covered by e2e.
12+
13+
tests:
14+
- it: renders the -csi HelmRelease when the tenant has etcd
15+
release:
16+
name: kubernetes-test
17+
namespace: tenant-test
18+
asserts:
19+
- hasDocuments:
20+
count: 1
21+
- isKind:
22+
of: HelmRelease
23+
- equal:
24+
path: metadata.name
25+
value: kubernetes-test-csi
26+
27+
- it: fallback yields replicated (default) and the kubevirt alias
28+
release:
29+
name: kubernetes-test
30+
namespace: tenant-test
31+
asserts:
32+
- equal:
33+
path: spec.values.storageClasses.replicated.infraStorageClass
34+
value: replicated
35+
- equal:
36+
path: spec.values.storageClasses.replicated.default
37+
value: true
38+
- equal:
39+
path: spec.values.storageClasses.kubevirt.infraStorageClass
40+
value: replicated
41+
- equal:
42+
path: spec.values.storageClasses.kubevirt.default
43+
value: false
44+
45+
- it: kubevirt alias tracks a custom storageClass selector, replicated stays default
46+
release:
47+
name: kubernetes-test
48+
namespace: tenant-test
49+
set:
50+
storageClass: fast
51+
asserts:
52+
- equal:
53+
path: spec.values.storageClasses.kubevirt.infraStorageClass
54+
value: fast
55+
- equal:
56+
path: spec.values.storageClasses.replicated.default
57+
value: true
58+
- equal:
59+
path: spec.values.storageClasses.kubevirt.default
60+
value: false

‎packages/apps/kubernetes/values.schema.json‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -3,7 +3,7 @@
33
"type": "object",
44
"properties": {
55
"storageClass": {
6-
"description": "StorageClass used to store the data.",
6+
"description": "Default StorageClass inside the tenant cluster. Remote-accessible LINSTOR classes are auto-propagated to the tenant under the same name.",
77
"type": "string",
88
"default": "replicated",
99
"x-kubernetes-validations": [

‎packages/apps/kubernetes/values.yaml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22
## @section Common Parameters
33
##
44

5-
## @param {string} storageClass - StorageClass used to store the data.
5+
## @param {string} storageClass - Default StorageClass inside the tenant cluster. Remote-accessible LINSTOR classes are auto-propagated to the tenant under the same name.
66
## @immutable
77
## @x-cozystack-options {source: storageclass}
88
storageClass: replicated

‎packages/system/kubernetes-rd/cozyrds/kubernetes.yaml‎

Lines changed: 1 addition & 1 deletion
Large diffs are not rendered by default.
Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,4 @@
1+
.PHONY: test
2+
3+
test:
4+
helm unittest .

0 commit comments

Comments
 (0)