diff --git a/api/apps/v1alpha1/kubernetes/types.go b/api/apps/v1alpha1/kubernetes/types.go index ea4a2b64..88a3c97e 100644 --- a/api/apps/v1alpha1/kubernetes/types.go +++ b/api/apps/v1alpha1/kubernetes/types.go @@ -69,6 +69,9 @@ type Addons struct { // NVIDIA GPU Operator. // +kubebuilder:default:={} GpuOperator GPUOperatorAddon `json:"gpuOperator"` + // HAMi GPU virtualization middleware. + // +kubebuilder:default:={} + Hami HAMiAddon `json:"hami"` // Ingress-NGINX controller. // +kubebuilder:default:={} IngressNginx IngressNginxAddon `json:"ingressNginx"` @@ -160,6 +163,15 @@ type GatewayAPIAddon struct { Enabled bool `json:"enabled"` } +type HAMiAddon struct { + // Enable HAMi (requires GPU Operator). + // +kubebuilder:default:=false + Enabled bool `json:"enabled"` + // Custom Helm values overrides. + // +kubebuilder:default:={} + ValuesOverride k8sRuntime.RawExtension `json:"valuesOverride"` +} + type Images struct { // Image used by the wait-for-kubeconfig init container. Empty falls back to images/busybox.tag. // +kubebuilder:default:="" diff --git a/api/apps/v1alpha1/kubernetes/zz_generated.deepcopy.go b/api/apps/v1alpha1/kubernetes/zz_generated.deepcopy.go index 2cbe3ba1..e021fbaa 100644 --- a/api/apps/v1alpha1/kubernetes/zz_generated.deepcopy.go +++ b/api/apps/v1alpha1/kubernetes/zz_generated.deepcopy.go @@ -49,6 +49,7 @@ func (in *Addons) DeepCopyInto(out *Addons) { in.Fluxcd.DeepCopyInto(&out.Fluxcd) out.GatewayAPI = in.GatewayAPI in.GpuOperator.DeepCopyInto(&out.GpuOperator) + in.Hami.DeepCopyInto(&out.Hami) in.IngressNginx.DeepCopyInto(&out.IngressNginx) in.MonitoringAgents.DeepCopyInto(&out.MonitoringAgents) in.Velero.DeepCopyInto(&out.Velero) @@ -135,6 +136,7 @@ func (in *ConfigSpec) DeepCopyInto(out *ConfigSpec) { } in.Addons.DeepCopyInto(&out.Addons) in.ControlPlane.DeepCopyInto(&out.ControlPlane) + out.Images = in.Images } // DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new ConfigSpec. @@ -260,6 +262,37 @@ func (in *GatewayAPIAddon) DeepCopy() *GatewayAPIAddon { return out } +// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil. +func (in *HAMiAddon) DeepCopyInto(out *HAMiAddon) { + *out = *in + in.ValuesOverride.DeepCopyInto(&out.ValuesOverride) +} + +// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new HAMiAddon. +func (in *HAMiAddon) DeepCopy() *HAMiAddon { + if in == nil { + return nil + } + out := new(HAMiAddon) + in.DeepCopyInto(out) + return out +} + +// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil. +func (in *Images) DeepCopyInto(out *Images) { + *out = *in +} + +// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new Images. +func (in *Images) DeepCopy() *Images { + if in == nil { + return nil + } + out := new(Images) + in.DeepCopyInto(out) + return out +} + // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil. func (in *IngressNginxAddon) DeepCopyInto(out *IngressNginxAddon) { *out = *in diff --git a/packages/apps/kubernetes/README.md b/packages/apps/kubernetes/README.md index ecbbc31b..bed1117a 100644 --- a/packages/apps/kubernetes/README.md +++ b/packages/apps/kubernetes/README.md @@ -128,6 +128,9 @@ See the reference for components utilized in this service: | `addons.gpuOperator` | NVIDIA GPU Operator. | `object` | `{}` | | `addons.gpuOperator.enabled` | Enable GPU Operator. | `bool` | `false` | | `addons.gpuOperator.valuesOverride` | Custom Helm values overrides. | `object` | `{}` | +| `addons.hami` | HAMi GPU virtualization middleware. | `object` | `{}` | +| `addons.hami.enabled` | Enable HAMi (requires GPU Operator). | `bool` | `false` | +| `addons.hami.valuesOverride` | Custom Helm values overrides. | `object` | `{}` | | `addons.fluxcd` | FluxCD GitOps operator. | `object` | `{}` | | `addons.fluxcd.enabled` | Enable FluxCD. | `bool` | `false` | | `addons.fluxcd.valuesOverride` | Custom Helm values overrides. | `object` | `{}` | diff --git a/packages/apps/kubernetes/templates/helmreleases/gpu-operator.yaml b/packages/apps/kubernetes/templates/helmreleases/gpu-operator.yaml index e243bfad..655dc868 100644 --- a/packages/apps/kubernetes/templates/helmreleases/gpu-operator.yaml +++ b/packages/apps/kubernetes/templates/helmreleases/gpu-operator.yaml @@ -1,3 +1,11 @@ +{{- define "cozystack.defaultGpuOperatorValues" -}} +{{- if .Values.addons.hami.enabled }} +gpu-operator: + devicePlugin: + enabled: false +{{- end }} +{{- end }} + {{- if and .Values.addons.gpuOperator.enabled .Values._namespace.etcd }} apiVersion: helm.toolkit.fluxcd.io/v2 kind: HelmRelease @@ -29,9 +37,12 @@ spec: force: true remediation: retries: -1 - {{- with .Values.addons.gpuOperator.valuesOverride }} + {{- $defaults := fromYaml (include "cozystack.defaultGpuOperatorValues" .) }} + {{- $overrides := deepCopy (default (dict) .Values.addons.gpuOperator.valuesOverride) }} + {{- $merged := mergeOverwrite (default (dict) $defaults) $overrides }} + {{- if $merged }} values: - {{- toYaml . | nindent 4 }} + {{- toYaml $merged | nindent 4 }} {{- end }} dependsOn: diff --git a/packages/apps/kubernetes/templates/helmreleases/hami.yaml b/packages/apps/kubernetes/templates/helmreleases/hami.yaml new file mode 100644 index 00000000..f1538c7c --- /dev/null +++ b/packages/apps/kubernetes/templates/helmreleases/hami.yaml @@ -0,0 +1,49 @@ +{{- if and .Values.addons.hami.enabled .Values._namespace.etcd }} +{{- if not .Values.addons.gpuOperator.enabled }} +{{- fail "addons.hami requires addons.gpuOperator to be enabled" }} +{{- end }} +apiVersion: helm.toolkit.fluxcd.io/v2 +kind: HelmRelease +metadata: + name: {{ .Release.Name }}-hami + labels: + cozystack.io/repository: system + cozystack.io/target-cluster-name: {{ .Release.Name }} + sharding.fluxcd.io/key: tenants +spec: + releaseName: hami + chartRef: + kind: ExternalArtifact + name: cozystack-kubernetes-application-kubevirt-kubernetes-hami + namespace: cozy-system + kubeConfig: + secretRef: + name: {{ .Release.Name }}-admin-kubeconfig + key: super-admin.svc + targetNamespace: cozy-hami + storageNamespace: cozy-hami + interval: 5m + timeout: 10m + install: + createNamespace: true + remediation: + retries: -1 + upgrade: + force: true + remediation: + retries: -1 + {{- with .Values.addons.hami.valuesOverride }} + values: + {{- toYaml . | nindent 4 }} + {{- end }} + + dependsOn: + {{- if lookup "helm.toolkit.fluxcd.io/v2" "HelmRelease" .Release.Namespace .Release.Name }} + - name: {{ .Release.Name }} + namespace: {{ .Release.Namespace }} + {{- end }} + - name: {{ .Release.Name }}-cilium + namespace: {{ .Release.Namespace }} + - name: {{ .Release.Name }}-gpu-operator + namespace: {{ .Release.Namespace }} +{{- end }} diff --git a/packages/apps/kubernetes/tests/gpu_operator_hami_test.yaml b/packages/apps/kubernetes/tests/gpu_operator_hami_test.yaml new file mode 100644 index 00000000..44528470 --- /dev/null +++ b/packages/apps/kubernetes/tests/gpu_operator_hami_test.yaml @@ -0,0 +1,99 @@ +suite: GPU Operator HelmRelease HAMi integration tests +templates: + - templates/helmreleases/gpu-operator.yaml +values: + - values-ci.yaml +tests: + - it: should disable devicePlugin when hami is enabled + set: + addons: + gpuOperator: + enabled: true + valuesOverride: {} + hami: + enabled: true + valuesOverride: {} + asserts: + - equal: + path: spec.values.gpu-operator.devicePlugin.enabled + value: false + + - it: should not have values when hami is disabled and no overrides + set: + addons: + gpuOperator: + enabled: true + valuesOverride: {} + hami: + enabled: false + valuesOverride: {} + asserts: + - notExists: + path: spec.values + + - it: should apply hami defaults when valuesOverride key is omitted + set: + addons: + gpuOperator: + enabled: true + hami: + enabled: true + asserts: + - equal: + path: spec.values.gpu-operator.devicePlugin.enabled + value: false + + - it: should allow user overrides to merge with hami defaults + set: + addons: + gpuOperator: + enabled: true + valuesOverride: + gpu-operator: + driver: + enabled: false + hami: + enabled: true + valuesOverride: {} + asserts: + - equal: + path: spec.values.gpu-operator.devicePlugin.enabled + value: false + - equal: + path: spec.values.gpu-operator.driver.enabled + value: false + + - it: should let user explicitly override devicePlugin.enabled to true with hami enabled + set: + addons: + gpuOperator: + enabled: true + valuesOverride: + gpu-operator: + devicePlugin: + enabled: true + driver: + enabled: false + hami: + enabled: true + valuesOverride: {} + asserts: + - equal: + path: spec.values.gpu-operator.devicePlugin.enabled + value: true + - equal: + path: spec.values.gpu-operator.driver.enabled + value: false + + - it: should not render when gpuOperator is disabled + set: + addons: + gpuOperator: + enabled: false + valuesOverride: {} + hami: + enabled: false + valuesOverride: {} + asserts: + - hasDocuments: + count: 0 diff --git a/packages/apps/kubernetes/tests/hami_test.yaml b/packages/apps/kubernetes/tests/hami_test.yaml new file mode 100644 index 00000000..27f7c75b --- /dev/null +++ b/packages/apps/kubernetes/tests/hami_test.yaml @@ -0,0 +1,153 @@ +suite: HAMi HelmRelease tests +templates: + - templates/helmreleases/hami.yaml +values: + - values-ci.yaml +tests: + - it: should not render when hami is disabled + set: + addons: + hami: + enabled: false + valuesOverride: {} + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - hasDocuments: + count: 0 + + - it: should render HelmRelease when hami is enabled + set: + addons: + hami: + enabled: true + valuesOverride: {} + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - hasDocuments: + count: 1 + - isKind: + of: HelmRelease + + - it: should fail when gpuOperator is not enabled + set: + addons: + hami: + enabled: true + valuesOverride: {} + gpuOperator: + enabled: false + valuesOverride: {} + asserts: + - failedTemplate: + errorMessage: "addons.hami requires addons.gpuOperator to be enabled" + + - it: should have correct metadata labels + set: + addons: + hami: + enabled: true + valuesOverride: {} + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - equal: + path: metadata.labels["cozystack.io/repository"] + value: system + - equal: + path: metadata.labels["sharding.fluxcd.io/key"] + value: tenants + + - it: should use ExternalArtifact chartRef + set: + addons: + hami: + enabled: true + valuesOverride: {} + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - equal: + path: spec.chartRef.kind + value: ExternalArtifact + - equal: + path: spec.chartRef.namespace + value: cozy-system + + - it: should target cozy-hami namespace + set: + addons: + hami: + enabled: true + valuesOverride: {} + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - equal: + path: spec.targetNamespace + value: cozy-hami + - equal: + path: spec.storageNamespace + value: cozy-hami + + - it: should depend on gpu-operator and cilium + release: + name: test + namespace: test-ns + set: + addons: + hami: + enabled: true + valuesOverride: {} + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - contains: + path: spec.dependsOn + content: + name: test-cilium + namespace: test-ns + - contains: + path: spec.dependsOn + content: + name: test-gpu-operator + namespace: test-ns + + - it: should not render spec.values when valuesOverride is empty + set: + addons: + hami: + enabled: true + valuesOverride: {} + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - hasDocuments: + count: 1 + - notExists: + path: spec.values + + - it: should pass through valuesOverride + set: + addons: + hami: + enabled: true + valuesOverride: + hami: + devicePlugin: + deviceSplitCount: 5 + gpuOperator: + enabled: true + valuesOverride: {} + asserts: + - equal: + path: spec.values.hami.devicePlugin.deviceSplitCount + value: 5 diff --git a/packages/apps/kubernetes/values.schema.json b/packages/apps/kubernetes/values.schema.json index 81e9e2b8..a43deaec 100644 --- a/packages/apps/kubernetes/values.schema.json +++ b/packages/apps/kubernetes/values.schema.json @@ -149,6 +149,7 @@ "fluxcd", "gatewayAPI", "gpuOperator", + "hami", "ingressNginx", "monitoringAgents", "velero", @@ -268,6 +269,28 @@ } } }, + "hami": { + "description": "HAMi GPU virtualization middleware.", + "type": "object", + "default": {}, + "required": [ + "enabled", + "valuesOverride" + ], + "properties": { + "enabled": { + "description": "Enable HAMi (requires GPU Operator).", + "type": "boolean", + "default": false + }, + "valuesOverride": { + "description": "Custom Helm values overrides.", + "type": "object", + "default": {}, + "x-kubernetes-preserve-unknown-fields": true + } + } + }, "ingressNginx": { "description": "Ingress-NGINX controller.", "type": "object", diff --git a/packages/apps/kubernetes/values.yaml b/packages/apps/kubernetes/values.yaml index 68ba07b0..d609476c 100644 --- a/packages/apps/kubernetes/values.yaml +++ b/packages/apps/kubernetes/values.yaml @@ -94,6 +94,10 @@ host: "" ## @field {bool} enabled - Enable FluxCD. ## @field {object} valuesOverride - Custom Helm values overrides. +## @typedef {struct} HAMiAddon - HAMi GPU virtualization middleware. +## @field {bool} enabled - Enable HAMi (requires GPU Operator). +## @field {object} valuesOverride - Custom Helm values overrides. + ## @typedef {struct} MonitoringAgentsAddon - Monitoring agents (Fluent Bit, VMAgents). ## @field {bool} enabled - Enable monitoring agents. ## @field {object} valuesOverride - Custom Helm values overrides. @@ -114,6 +118,7 @@ host: "" ## @field {GatewayAPIAddon} gatewayAPI - Gateway API addon. ## @field {IngressNginxAddon} ingressNginx - Ingress-NGINX controller. ## @field {GPUOperatorAddon} gpuOperator - NVIDIA GPU Operator. +## @field {HAMiAddon} hami - HAMi GPU virtualization middleware. ## @field {FluxCDAddon} fluxcd - FluxCD GitOps operator. ## @field {MonitoringAgentsAddon} monitoringAgents - Monitoring agents. ## @field {VerticalPodAutoscalerAddon} verticalPodAutoscaler - Vertical Pod Autoscaler. @@ -137,6 +142,9 @@ addons: gpuOperator: enabled: false valuesOverride: {} + hami: + enabled: false + valuesOverride: {} fluxcd: enabled: false valuesOverride: {} diff --git a/packages/core/platform/sources/hami.yaml b/packages/core/platform/sources/hami.yaml new file mode 100644 index 00000000..0184e405 --- /dev/null +++ b/packages/core/platform/sources/hami.yaml @@ -0,0 +1,24 @@ +--- +apiVersion: cozystack.io/v1alpha1 +kind: PackageSource +metadata: + name: cozystack.hami +spec: + sourceRef: + kind: OCIRepository + name: cozystack-packages + namespace: cozy-system + path: / + variants: + - name: default + dependsOn: + - cozystack.gpu-operator + components: + - name: hami + path: system/hami + valuesFiles: + - values.yaml + install: + privileged: true + namespace: cozy-hami + releaseName: hami diff --git a/packages/core/platform/sources/kubernetes-application.yaml b/packages/core/platform/sources/kubernetes-application.yaml index 9787cb19..088383ad 100644 --- a/packages/core/platform/sources/kubernetes-application.yaml +++ b/packages/core/platform/sources/kubernetes-application.yaml @@ -52,6 +52,8 @@ spec: path: system/cilium - name: kubernetes-gpu-operator path: system/gpu-operator + - name: kubernetes-hami + path: system/hami - name: kubernetes-vertical-pod-autoscaler path: system/vertical-pod-autoscaler - name: kubernetes-prometheus-operator-crds diff --git a/packages/core/platform/templates/bundles/iaas.yaml b/packages/core/platform/templates/bundles/iaas.yaml index ced0a322..66ec98b2 100644 --- a/packages/core/platform/templates/bundles/iaas.yaml +++ b/packages/core/platform/templates/bundles/iaas.yaml @@ -10,6 +10,7 @@ {{include "cozystack.platform.package.default" (list "cozystack.kubevirt-cdi" $) }} {{include "cozystack.platform.package.optional.default" (list "cozystack.vm-default-images" $) }} {{include "cozystack.platform.package.optional.default" (list "cozystack.gpu-operator" $) }} +{{include "cozystack.platform.package.optional.default" (list "cozystack.hami" $) }} {{include "cozystack.platform.package.default" (list "cozystack.kamaji" $) }} {{include "cozystack.platform.package.default" (list "cozystack.capi-operator" $) }} {{include "cozystack.platform.package.default" (list "cozystack.capi-provider-bootstrap-kubeadm" $) }} diff --git a/packages/system/hami/Chart.yaml b/packages/system/hami/Chart.yaml new file mode 100644 index 00000000..3fcf4c5d --- /dev/null +++ b/packages/system/hami/Chart.yaml @@ -0,0 +1,3 @@ +apiVersion: v2 +name: cozy-hami +version: 0.0.0 # Placeholder, the actual version will be automatically set during the build process diff --git a/packages/system/hami/Makefile b/packages/system/hami/Makefile new file mode 100644 index 00000000..83663a66 --- /dev/null +++ b/packages/system/hami/Makefile @@ -0,0 +1,43 @@ +export NAME=hami +export NAMESPACE=cozy-$(NAME) + +include ../../../hack/common-envs.mk +include ../../../hack/package.mk + +# When bumping the HAMi version, run `make update` and then review +# the resulting diff in `charts/hami/`. The recipe below reproduces the +# top-level vendoring overrides automatically: +# +# 1. Removes the broken hami-dra subchart. Upstream's NVIDIA DRA driver +# path requires kubelet DRA support that cozystack does not enable +# and has no upstream fix tracked. See commit 3c5521e. +# 2. Empties Chart.yaml dependencies and drops Chart.lock so Helm does +# not try to re-pull hami-dra at build time. See commit 2734dc0. +# 3. Strips dra/hami-dra/podSecurityPolicy blocks from the upstream +# values.yaml since the corresponding code paths are gone. PSP is +# removed from Kubernetes 1.25+ and is unused by cozystack. +# +# Template-level patches are NOT reproduced automatically: +# +# * Scheduler templates have `{{- if .Values.dra.enabled }}` blocks +# that need to be removed because the dra value is gone (commit +# 2734dc0 stripped them). +# * device-plugin/monitorservice.yaml uses `indent` with leading +# whitespace; it must be rewritten to `nindent` for the labels block +# to render correctly when devicePlugin.service.labels is set +# (commit 3685254). +# +# After `make update`, run `git diff -- charts/hami/templates/` and +# replay these template patches against the new upstream version, then +# verify with `helm unittest`. If upstream restructured the affected +# files, the patches may need to be redesigned rather than reapplied. + +update: + rm -rf charts + helm repo add hami-charts https://project-hami.github.io/HAMi/ + helm repo update hami-charts + helm pull hami-charts/hami --untar --untardir charts + rm -rf charts/hami/charts/hami-dra + yq --inplace '.dependencies = []' charts/hami/Chart.yaml + rm -f charts/hami/Chart.lock + yq --inplace 'del(.dra) | del(.["hami-dra"]) | del(.podSecurityPolicy)' charts/hami/values.yaml diff --git a/packages/system/hami/README.md b/packages/system/hami/README.md new file mode 100644 index 00000000..68669ac7 --- /dev/null +++ b/packages/system/hami/README.md @@ -0,0 +1,82 @@ +# HAMi — GPU Virtualization Middleware + +[HAMi](https://github.com/Project-HAMi/HAMi) (Heterogeneous AI Computing Virtualization Middleware) is a CNCF Sandbox project that enables fractional GPU sharing in Kubernetes. It allows workloads to request specific amounts of GPU memory and compute cores instead of claiming entire GPUs. + +## Architecture + +HAMi consists of four components: + +- **MutatingWebhook** — intercepts pod creation, injects `schedulerName: hami-scheduler` +- **Scheduler Extender** — extends kube-scheduler with GPU-aware Filter and Bind logic +- **Device Plugin** (DaemonSet) — registers vGPU resources via the Kubernetes Device Plugin API +- **HAMi-core** (`libvgpu.so`) — `LD_PRELOAD` library injected into workload containers, intercepts CUDA API calls to enforce memory and compute isolation + +## Prerequisites + +- GPU Operator must be enabled (`addons.gpuOperator.enabled: true`) +- NVIDIA driver >= 440 on host nodes +- nvidia-container-toolkit configured as the default container runtime +- GPU nodes labeled with `gpu=on` + +## Known Limitations + +### glibc < 2.34 requirement for workload containers + +HAMi-core uses `LD_PRELOAD` to intercept `dlsym()` for CUDA symbol resolution. The fallback code path relies on `_dl_sym`, a private glibc internal symbol that was removed in glibc 2.34 when libdl and libpthread were merged into libc.so. + +**This limitation affects workload containers only**, not the host OS or HAMi's own components. + +| Distribution | glibc | Result | +| --------------- | ----- | -------------------------------------------- | +| Ubuntu 18.04 | 2.27 | Full isolation (memory + compute) | +| Ubuntu 20.04 | 2.31 | Full isolation (memory + compute) | +| Ubuntu 22.04 | 2.35 | Memory isolation works, compute breaks | +| Ubuntu 24.04 | 2.39 | Both memory and compute isolation break | +| Alpine (musl) | N/A | Completely incompatible (`dlvsym` absent) | + +Most modern ML/AI base images (CUDA 12.x, PyTorch 2.x, TensorFlow 2.x) use Ubuntu 22.04+ with glibc >= 2.35, which means compute isolation will not work with these images until the upstream fix is merged. + +**Upstream tracking issues:** + +- [HAMi-core#174](https://github.com/Project-HAMi/HAMi-core/issues/174) — `_dl_sym` removal in glibc 2.34 breaks HAMi-core's CUDA symbol resolution at the symbol level +- [HAMi#1190](https://github.com/Project-HAMi/HAMi/issues/1190) — maintainer thread confirming the empirical per-glibc-version isolation behavior shown in the table above + +### musl libc (Alpine) incompatibility + +HAMi-core is completely incompatible with musl libc. The `dlvsym()` function used by HAMi-core is a glibc extension not available in musl. Only glibc-based container images (Debian, Ubuntu, RHEL, etc.) can use HAMi GPU isolation. + +## Usage + +Enable HAMi in your tenant Kubernetes cluster values: + +```yaml +addons: + gpuOperator: + enabled: true + hami: + enabled: true +``` + +When HAMi is enabled, GPU Operator's built-in device plugin is automatically disabled to avoid conflicts. This default is preserved by setting `addons.gpuOperator.valuesOverride.gpu-operator.devicePlugin.enabled: false`; advanced topologies that partition GPU pools (e.g. some nodes use HAMi while others run the standard NVIDIA device plugin via node selectors) can re-enable it explicitly through `valuesOverride`. + +### Requesting fractional GPU resources + +```yaml +resources: + limits: + nvidia.com/gpu: 1 + nvidia.com/gpumem: 3000 # 3000 MB of GPU memory + nvidia.com/gpucores: 30 # 30% of GPU compute cores +``` + +## Parameters + +Default values shown below are inherited from the upstream HAMi chart and may change with upstream updates. + +| Name | Description | Default | +| --- | --- | --- | +| `hami.devicePlugin.runtimeClassName` | RuntimeClass for device plugin pods | `nvidia` | +| `hami.devicePlugin.deviceSplitCount` | Max virtual GPUs per physical GPU | `10` | +| `hami.devicePlugin.deviceMemoryScaling` | Memory overcommit factor (> 1.0 enables overcommit) | `1` | +| `hami.scheduler.defaultSchedulerPolicy.nodeSchedulerPolicy` | Node packing strategy (`binpack` or `spread`) | `binpack` | +| `hami.scheduler.defaultSchedulerPolicy.gpuSchedulerPolicy` | GPU packing strategy (`binpack` or `spread`) | `spread` | diff --git a/packages/system/hami/charts/hami/Chart.yaml b/packages/system/hami/charts/hami/Chart.yaml new file mode 100644 index 00000000..55f32ab6 --- /dev/null +++ b/packages/system/hami/charts/hami/Chart.yaml @@ -0,0 +1,18 @@ +apiVersion: v2 +appVersion: 2.8.1 +dependencies: [] +description: Heterogeneous AI Computing Virtualization Middleware +keywords: +- vgpu +- gpu +kubeVersion: '>= 1.18.0-0' +maintainers: +- email: archlitchi@gmail.com + name: limengxuan +- email: xiaozhang0210@hotmail.com + name: zhangxiao +name: hami +sources: +- https://github.com/Project-HAMi/HAMi +type: application +version: 2.8.1 diff --git a/packages/system/hami/charts/hami/README.md b/packages/system/hami/charts/hami/README.md new file mode 100644 index 00000000..e8217290 --- /dev/null +++ b/packages/system/hami/charts/hami/README.md @@ -0,0 +1,237 @@ +# HAMi Helm Chart Values Documentation + +This document provides detailed descriptions of all configurable values parameters for the HAMi Helm Chart. + +## Global Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `global.imageRegistry` | Global Docker image registry | `""` | +| `global.imagePullSecrets` | Global Docker image pull secrets | `[]` | +| `global.imageTag` | Image tag | `"v2.8.1"` | +| `global.gpuHookPath` | GPU Hook path | `/usr/local` | +| `global.labels` | Global labels | `{}` | +| `global.annotations` | Global annotations | `{}` | +| `global.managedNodeSelectorEnable` | Whether to enable managed node selector | `false` | +| `global.managedNodeSelector.usage` | Managed node selector usage | `"gpu"` | +| `nameOverride` | Name override | `""` | +| `fullnameOverride` | Full name override | `""` | +| `namespaceOverride` | Namespace override | `""` | + +## Resource Name Configuration + +### NVIDIA GPU Resources +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `resourceName` | GPU resource name | `"nvidia.com/gpu"` | +| `resourceMem` | GPU memory resource name | `"nvidia.com/gpumem"` | +| `resourceMemPercentage` | GPU memory percentage resource name | `"nvidia.com/gpumem-percentage"` | +| `resourceCores` | GPU core resource name | `"nvidia.com/gpucores"` | +| `resourcePriority` | GPU priority resource name | `"nvidia.com/priority"` | + +### Cambricon MLU Resources +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `mluResourceName` | MLU resource name | `"cambricon.com/vmlu"` | +| `mluResourceMem` | MLU memory resource name | `"cambricon.com/mlu.smlu.vmemory"` | +| `mluResourceCores` | MLU core resource name | `"cambricon.com/mlu.smlu.vcore"` | + +### Hygon DCU Resources +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `dcuResourceName` | DCU resource name | `"hygon.com/dcunum"` | +| `dcuResourceMem` | DCU memory resource name | `"hygon.com/dcumem"` | +| `dcuResourceCores` | DCU core resource name | `"hygon.com/dcucores"` | + +### Metax GPU Resources +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `metaxResourceName` | GPU resource name | `"metax-tech.com/sgpu"` | +| `metaxResourceCore` | GPU core resource name | `"metax-tech.com/vcore"` | +| `metaxResourceMem` | GPU memory resource name | `"metax-tech.com/vmemory"` | +| `metaxsGPUTopologyAware` | GPU topology awareness | `"false"` | + +### Enflame GCU Resources +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `enflameResourceNameVGCU` | vGCU resource name | `"enflame.com/vgcu"` | +| `enflameResourceNameVGCUPercentage` | vGCU percentage resource name | `"enflame.com/vgcu-percentage"` | + +### Kunlunxin XPU Resources +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `kunlunResourceName` | XPU resource name | `"kunlunxin.com/xpu"` | + +## Scheduler Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `schedulerName` | Scheduler name | `"hami-scheduler"` | +| `scheduler.nodeName` | Define node name, scheduler will schedule to this node | `""` | +| `scheduler.overwriteEnv` | Whether to overwrite environment variables | `"false"` | +| `scheduler.defaultSchedulerPolicy.nodeSchedulerPolicy` | Node scheduler policy | `binpack` | +| `scheduler.defaultSchedulerPolicy.gpuSchedulerPolicy` | GPU scheduler policy | `spread` | +| `scheduler.metricsBindAddress` | Metrics bind address | `":9395"` | +| `scheduler.forceOverwriteDefaultScheduler` | Whether to force overwrite default scheduler | `true` | +| `scheduler.livenessProbe` | Whether to enable liveness probe | `false` | +| `scheduler.leaderElect` | Whether to enable leader election | `true` | +| `scheduler.replicas` | Number of replicas | `1` | + +### Kube Scheduler Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `scheduler.kubeScheduler.enabled` | Whether to run kube-scheduler container in scheduler pod | `true` | +| `scheduler.kubeScheduler.image.registry` | Kube scheduler image registry | `"registry.cn-hangzhou.aliyuncs.com"` | +| `scheduler.kubeScheduler.image.repository` | Kube scheduler image repository | `"google_containers/kube-scheduler"` | +| `scheduler.kubeScheduler.image.tag` | Kube scheduler image tag | `""` | +| `scheduler.kubeScheduler.image.pullPolicy` | Kube scheduler image pull policy | `IfNotPresent` | +| `scheduler.kubeScheduler.image.pullSecrets` | Kube scheduler image pull secrets | `[]` | +| `scheduler.kubeScheduler.extraNewArgs` | Extra new arguments | `["--config=/config/config.yaml", "-v=4"]` | +| `scheduler.kubeScheduler.extraArgs` | Extra arguments | `["--policy-config-file=/config/config.json", "-v=4"]` | + +### Extender Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `scheduler.extender.image.registry` | Scheduler extender image registry | `"docker.io"` | +| `scheduler.extender.image.repository` | Scheduler extender image repository | `"projecthami/hami"` | +| `scheduler.extender.image.tag` | Scheduler extender image tag | `""` | +| `scheduler.extender.image.pullPolicy` | Scheduler extender image pull policy | `IfNotPresent` | +| `scheduler.extender.image.pullSecrets` | Scheduler extender image pull secrets | `[]` | +| `scheduler.extender.extraArgs` | Scheduler extender extra arguments | `["--debug", "-v=4"]` | + +### Admission Webhook Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `scheduler.admissionWebhook.enabled` | Whether to enable admission webhook | `true` | +| `scheduler.admissionWebhook.customURL.enabled` | Whether to enable custom URL | `false` | +| `scheduler.admissionWebhook.customURL.host` | Custom URL host | `127.0.0.1` | +| `scheduler.admissionWebhook.customURL.port` | Custom URL port | `31998` | +| `scheduler.admissionWebhook.customURL.path` | Custom URL path | `/webhook` | +| `scheduler.admissionWebhook.reinvocationPolicy` | Reinvocation policy | `Never` | +| `scheduler.admissionWebhook.failurePolicy` | Failure policy | `Ignore` | + +### TLS Certificate Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `scheduler.certManager.enabled` | Whether to use cert-manager to generate self-signed certificates | `false` | +| `scheduler.patch.enabled` | Whether to use kube-webhook-certgen to generate self-signed certificates | `true` | +| `scheduler.patch.image.registry` | Certgen image registry | `"docker.io"` | +| `scheduler.patch.image.repository` | Certgen image repository | `"jettech/kube-webhook-certgen"` | +| `scheduler.patch.image.tag` | Certgen image tag | `"v1.5.2"` | +| `scheduler.patch.image.pullPolicy` | Certgen image pull policy | `IfNotPresent` | +| `scheduler.patch.imageNew.registry` | New certgen image registry | `"docker.io"` | +| `scheduler.patch.imageNew.repository` | New certgen image repository | `"liangjw/kube-webhook-certgen"` | +| `scheduler.patch.imageNew.tag` | New certgen image tag | `"v1.1.1"` | + +### Scheduler Service Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `scheduler.service.type` | Service type | `NodePort` | +| `scheduler.service.httpPort` | HTTP port | `443` | +| `scheduler.service.schedulerPort` | Scheduler NodePort | `31998` | +| `scheduler.service.monitorPort` | Monitor port | `31993` | +| `scheduler.service.monitorTargetPort` | Monitor target port | `9395` | + +## Device Plugin Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devicePlugin.image.registry` | Device plugin image registry | `"docker.io"` | +| `devicePlugin.image.repository` | Device plugin image repository | `"projecthami/hami"` | +| `devicePlugin.image.tag` | Device plugin image tag | `""` | +| `devicePlugin.image.pullPolicy` | Device plugin image pull policy | `IfNotPresent` | +| `devicePlugin.image.pullSecrets` | Device plugin image pull secrets | `[]` | + +### Monitor Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devicePlugin.monitor.image.registry` | Monitor image registry | `"docker.io"` | +| `devicePlugin.monitor.image.repository` | Monitor image repository | `"projecthami/hami"` | +| `devicePlugin.monitor.image.tag` | Monitor image tag | `""` | +| `devicePlugin.monitor.image.pullPolicy` | Monitor image pull policy | `IfNotPresent` | +| `devicePlugin.monitor.image.pullSecrets` | Monitor image pull secrets | `[]` | +| `devicePlugin.monitor.ctrPath` | Container path | `/usr/local/vgpu/containers` | +| `devicePlugin.monitor.extraArgs` | Monitor extra arguments | `["-v=4"]` | +| `devicePlugin.monitor.extraEnvs` | Monitor extra environments | `{}` | + +### Device Plugin Other Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devicePlugin.deviceSplitCount` | Integer type, default value: 10. Maximum number of tasks assigned to a single GPU device | `10` | +| `devicePlugin.deviceMemoryScaling` | Device memory scaling ratio | `1` | +| `devicePlugin.deviceCoreScaling` | Device core scaling ratio | `1` | +| `devicePlugin.runtimeClassName` | Runtime class name | `""` | +| `devicePlugin.createRuntimeClass` | Whether to create runtime class | `false` | +| `devicePlugin.migStrategy` | String type, "none" means ignore MIG functionality, "mixed" means allocate MIG devices through independent resources | `"none"` | +| `devicePlugin.disablecorelimit` | String type, "true" means disable core limit, "false" means enable core limit | `"false"` | +| `devicePlugin.passDeviceSpecsEnabled` | Whether to enable passing device specs | `false` | +| `devicePlugin.extraArgs` | Device plugin extra arguments | `["-v=4"]` | +| `devicePlugin.nodeConfiguration.config` | Node configuration for device plugin by json | An example of default configuration. | +| `devicePlugin.nodeConfiguration.externalConfigName` | Node configuration for device plugin by external congimap | `""` | +| `devicePlugin.extraEnvs` | Device plugin extra environments | `{}` | + +### Device Plugin Service Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devicePlugin.service.type` | Service type | `NodePort` | +| `devicePlugin.service.httpPort` | HTTP port | `31992` | + +### Device Plugin Deployment Configuration + +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devicePlugin.pluginPath` | Plugin path | `/var/lib/kubelet/device-plugins` | +| `devicePlugin.libPath` | Library path | `/usr/local/vgpu` | +| `devicePlugin.nvidiaNodeSelector` | NVIDIA node selector | `{"gpu": "on"}` | +| `devicePlugin.updateStrategy.type` | Update strategy type | `RollingUpdate` | +| `devicePlugin.updateStrategy.rollingUpdate.maxUnavailable` | Maximum unavailable count | `1` | + +## Device Configuration + +### AWS Neuron +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devices.awsneuron.customresources` | Custom resources | `["aws.amazon.com/neuron", "aws.amazon.com/neuroncore"]` | + +### Kunlunxin +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devices.kunlun.enabled` | Whether to enable | `true` | +| `devices.kunlun.customresources` | Custom resources | `["kunlunxin.com/xpu"]` | + +### Mthreads +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devices.mthreads.enabled` | Whether to enable | `true` | +| `devices.mthreads.customresources` | Custom resources | `["mthreads.com/vgpu"]` | + +### NVIDIA +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devices.nvidia.gpuCorePolicy` | GPU core policy | `default` | +| `devices.nvidia.libCudaLogLevel` | CUDA library log level | `1` | + +### Huawei Ascend +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devices.ascend.enabled` | Whether to enable | `false` | +| `devices.ascend.image` | Image | `""` | +| `devices.ascend.imagePullPolicy` | Image pull policy | `IfNotPresent` | +| `devices.ascend.extraArgs` | Extra arguments | `[]` | +| `devices.ascend.nodeSelector` | Node selector | `{"ascend": "on"}` | +| `devices.ascend.tolerations` | Tolerations | `[]` | +| `devices.ascend.customresources` | Custom resources | `["huawei.com/Ascend910A", "huawei.com/Ascend910A-memory", ...]` | + +### Iluvatar +| Parameter | Description | Default Value | +|-----------|-------------|---------------| +| `devices.iluvatar.enabled` | Whether to enable | `false` | +| `devices.iluvatar.customresources` | Custom resources | `["iluvatar.ai/BI-V150-vgpu", "iluvatar.ai/BI-V150.vMem","iluvatar.ai/BI-V150.vCore", ...]` | diff --git a/packages/system/hami/charts/hami/templates/NOTES.txt b/packages/system/hami/charts/hami/templates/NOTES.txt new file mode 100644 index 00000000..15bb1218 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/NOTES.txt @@ -0,0 +1,3 @@ +** Please be patient while the chart is being deployed ** +Resource name: {{ .Values.resourceName }} + diff --git a/packages/system/hami/charts/hami/templates/_commons.tpl b/packages/system/hami/charts/hami/templates/_commons.tpl new file mode 100644 index 00000000..b68018e4 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/_commons.tpl @@ -0,0 +1,49 @@ +{{/* +Return the proper image name +{{ include "common.images.image" ( dict "imageRoot" .Values.path.to.the.image "global" $) }} +*/}} +{{- define "common.images.image" -}} +{{- $registryName := .imageRoot.registry -}} +{{- $repositoryName := .imageRoot.repository -}} +{{- $tag := .imageRoot.tag | toString -}} +{{- if .global }} + {{- if .global.imageRegistry }} + {{- $registryName = .global.imageRegistry -}} + {{- end -}} +{{- end -}} +{{- if .tag }} + {{- $tag = .tag | toString -}} +{{- end -}} +{{- if $registryName }} +{{- printf "%s/%s:%s" $registryName $repositoryName $tag -}} +{{- else -}} +{{- printf "%s:%s" $repositoryName $tag -}} +{{- end -}} +{{- end -}} + +{{/* +Return the proper Docker Image Registry Secret Names (deprecated: use common.images.renderPullSecrets instead) +{{ include "common.images.pullSecrets" ( dict "images" (list .Values.path.to.the.image1, .Values.path.to.the.image2) "global" .Values.global) }} +*/}} +{{- define "common.images.pullSecrets" -}} + {{- $pullSecrets := list }} + + {{- if .global }} + {{- range .global.imagePullSecrets -}} + {{- $pullSecrets = append $pullSecrets . -}} + {{- end -}} + {{- end -}} + + {{- range .images -}} + {{- range .pullSecrets -}} + {{- $pullSecrets = append $pullSecrets . -}} + {{- end -}} + {{- end -}} + + {{- if (not (empty $pullSecrets)) }} +imagePullSecrets: + {{- range $pullSecrets }} + - name: {{ . }} + {{- end }} + {{- end }} +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/_helpers.tpl b/packages/system/hami/charts/hami/templates/_helpers.tpl new file mode 100644 index 00000000..ffcc61da --- /dev/null +++ b/packages/system/hami/charts/hami/templates/_helpers.tpl @@ -0,0 +1,163 @@ +{{/* +Expand the name of the chart. +*/}} +{{- define "hami-vgpu.name" -}} +{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" -}} +{{- end -}} + +{{/* +Create a default fully qualified app name. +We truncate at 63 chars because some Kubernetes name fields are limited to this (by the DNS naming spec). +If release name contains chart name it will be used as a full name. +*/}} +{{- define "hami-vgpu.fullname" -}} +{{- if .Values.fullnameOverride -}} +{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" -}} +{{- else }} +{{- $name := default .Chart.Name .Values.nameOverride -}} +{{- if contains $name .Release.Name }} +{{- .Release.Name | trunc 63 | trimSuffix "-" -}} +{{- else -}} +{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" -}} +{{- end -}} +{{- end -}} +{{- end -}} + +{{/* +Allow the release namespace to be overridden for multi-namespace deployments in combined charts +*/}} +{{- define "hami-vgpu.namespace" -}} + {{- if .Values.namespaceOverride -}} + {{- .Values.namespaceOverride -}} + {{- else -}} + {{- .Release.Namespace -}} + {{- end -}} +{{- end -}} + +{{/* +The app name for Scheduler +*/}} +{{- define "hami-vgpu.scheduler" -}} +{{- printf "%s-scheduler" ( include "hami-vgpu.fullname" . ) | trunc 63 | trimSuffix "-" -}} +{{- end -}} + +{{/* +The app name for DevicePlugin +*/}} +{{- define "hami-vgpu.device-plugin" -}} +{{- printf "%s-device-plugin" ( include "hami-vgpu.fullname" . ) | trunc 63 | trimSuffix "-" -}} +{{- end -}} + +{{/* + The app name for MockDevicePlugin + */}} +{{- define "hami-vgpu.mock-device-plugin" -}} +{{- printf "%s-mock-device-plugin" ( include "hami-vgpu.fullname" . ) | trunc 63 | trimSuffix "-" -}} +{{- end -}} + +{{/* +The tls secret name for Scheduler +*/}} +{{- define "hami-vgpu.scheduler.tls" -}} +{{- printf "%s-scheduler-tls" ( include "hami-vgpu.fullname" . ) | trunc 63 | trimSuffix "-" -}} +{{- end -}} + +{{/* +The webhook name +*/}} +{{- define "hami-vgpu.scheduler.webhook" -}} +{{- printf "%s-webhook" ( include "hami-vgpu.fullname" . ) | trunc 63 | trimSuffix "-" -}} +{{- end -}} + +{{/* +Create chart name and version as used by the chart label. +*/}} +{{- define "hami-vgpu.chart" -}} +{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" }} +{{- end }} + +{{/* +Common labels +*/}} +{{- define "hami-vgpu.labels" -}} +helm.sh/chart: {{ include "hami-vgpu.chart" . }} +{{ include "hami-vgpu.selectorLabels" . }} +{{- if .Chart.AppVersion }} +app.kubernetes.io/version: {{ .Chart.AppVersion | quote }} +{{- end }} +app.kubernetes.io/managed-by: {{ .Release.Service }} +{{- end }} + +{{/* +Selector labels +*/}} +{{- define "hami-vgpu.selectorLabels" -}} +app.kubernetes.io/name: {{ include "hami-vgpu.name" . }} +app.kubernetes.io/instance: {{ .Release.Name }} +{{- end }} + + +{{/* + Resolve the tag for kubeScheduler. +*/}} +{{- define "resolvedKubeSchedulerTag" -}} +{{- if .Values.scheduler.kubeScheduler.image.tag }} +{{- .Values.scheduler.kubeScheduler.image.tag | trim -}} +{{- else }} +{{- include "strippedKubeVersion" . | trim -}} +{{- end }} +{{- end }} + +{{/* + Return the stripped Kubernetes version string by removing extra parts after semantic version number. + v1.31.1+k3s1 -> v1.31.1 + v1.30.8-eks-2d5f260 -> v1.30.8 + v1.31.1 -> v1.31.1 +*/}} +{{- define "strippedKubeVersion" -}} +{{ regexReplaceAll "^(v[0-9]+\\.[0-9]+\\.[0-9]+)(.*)$" .Capabilities.KubeVersion.Version "$1" }} +{{- end -}} + +{{- define "hami.scheduler.kubeScheduler.image" -}} +{{ include "common.images.image" (dict "imageRoot" .Values.scheduler.kubeScheduler.image "global" .Values.global "tag" (include "resolvedKubeSchedulerTag" .)) }} +{{- end -}} + +{{- define "hami.scheduler.extender.image" -}} +{{ include "common.images.image" (dict "imageRoot" .Values.scheduler.extender.image "global" .Values.global "tag" .Values.global.imageTag) }} +{{- end -}} + +{{- define "hami.devicePlugin.image" -}} +{{ include "common.images.image" (dict "imageRoot" .Values.devicePlugin.image "global" .Values.global "tag" .Values.global.imageTag) }} +{{- end -}} + +{{- define "hami.mockDevicePlugin.image" -}} +{{ include "common.images.image" (dict "imageRoot" .Values.mockDevicePlugin.image "global" .Values.global "tag" .Values.mockDevicePlugin.tag) }} +{{- end -}} + +{{- define "hami.devicePlugin.monitor.image" -}} +{{ include "common.images.image" (dict "imageRoot" .Values.devicePlugin.monitor.image "global" .Values.global "tag" .Values.global.imageTag) }} +{{- end -}} + +{{- define "hami.scheduler.patch.image" -}} +{{ include "common.images.image" (dict "imageRoot" .Values.scheduler.patch.image "global" .Values.global) }} +{{- end -}} + +{{- define "hami.scheduler.patch.new.image" -}} +{{ include "common.images.image" (dict "imageRoot" .Values.scheduler.patch.imageNew "global" .Values.global) }} +{{- end -}} + +{{- define "hami.scheduler.extender.imagePullSecrets" -}} +{{ include "common.images.pullSecrets" (dict "images" (list .Values.scheduler.extender.image) "global" .Values.global) }} +{{- end -}} + +{{- define "hami.devicePlugin.imagePullSecrets" -}} +{{ include "common.images.pullSecrets" (dict "images" (list .Values.devicePlugin.image) "global" .Values.global) }} +{{- end -}} + +{{- define "hami.scheduler.patch.imagePullSecrets" -}} +{{ include "common.images.pullSecrets" (dict "images" (list .Values.scheduler.patch.image) "global" .Values.global) }} +{{- end -}} + +{{- define "hami.scheduler.patch.new.imagePullSecrets" -}} +{{ include "common.images.pullSecrets" (dict "images" (list .Values.scheduler.patch.imageNew) "global" .Values.global) }} +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/device-plugin/configmap.yaml b/packages/system/hami/charts/hami/templates/device-plugin/configmap.yaml new file mode 100644 index 00000000..74631e24 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/configmap.yaml @@ -0,0 +1,13 @@ +{{- if and .Values.devicePlugin.enabled (not .Values.devicePlugin.nodeConfiguration.externalConfigName) -}} +apiVersion: v1 +kind: ConfigMap +metadata: + name: {{ include "hami-vgpu.device-plugin" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-device-plugin + {{- include "hami-vgpu.labels" . | nindent 4 }} +data: + config.json: | +{{- .Values.devicePlugin.nodeConfiguration.config | nindent 4 }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/device-plugin/daemonsetmock.yaml b/packages/system/hami/charts/hami/templates/device-plugin/daemonsetmock.yaml new file mode 100644 index 00000000..a0dc4ff3 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/daemonsetmock.yaml @@ -0,0 +1,55 @@ +{{- if .Values.mockDevicePlugin.enabled }} +apiVersion: apps/v1 +kind: DaemonSet +metadata: + name: {{ include "hami-vgpu.mock-device-plugin" . }} + namespace: {{ include "hami-vgpu.namespace" . }} +spec: + selector: + matchLabels: + app.kubernetes.io/component: hami-mock-device-plugin + {{- include "hami-vgpu.selectorLabels" . | nindent 6 }} + template: + metadata: + annotations: + scheduler.alpha.kubernetes.io/critical-pod: "" + labels: + app.kubernetes.io/component: hami-mock-device-plugin + {{- include "hami-vgpu.selectorLabels" . | nindent 8 }} + spec: + serviceAccountName: {{ include "hami-vgpu.mock-device-plugin" . }} + tolerations: + - key: CriticalAddonsOnly + operator: Exists + containers: + - image: {{ include "hami.mockDevicePlugin.image" . }} + imagePullPolicy: {{ .Values.mockDevicePlugin.image.pullPolicy }} + name: hami-mock-dp-cntr + env: + - name: NODE_NAME + valueFrom: + fieldRef: + fieldPath: spec.nodeName + command: + - ./k8s-device-plugin + - -v=5 + - --device-config-file=/device-config.yaml + volumeMounts: + - name: dp + mountPath: /var/lib/kubelet/device-plugins + - name: sys + mountPath: /sys + - name: device-config + mountPath: /device-config.yaml + subPath: device-config.yaml + volumes: + - name: dp + hostPath: + path: /var/lib/kubelet/device-plugins + - name: sys + hostPath: + path: /sys + - name: device-config + configMap: + name: {{ include "hami-vgpu.scheduler" . }}-device +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/device-plugin/daemonsetnvidia.yaml b/packages/system/hami/charts/hami/templates/device-plugin/daemonsetnvidia.yaml new file mode 100644 index 00000000..1f4c24f3 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/daemonsetnvidia.yaml @@ -0,0 +1,262 @@ +{{- if .Values.devicePlugin.enabled }} +apiVersion: apps/v1 +kind: DaemonSet +metadata: + name: {{ include "hami-vgpu.device-plugin" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-device-plugin + {{- include "hami-vgpu.labels" . | nindent 4 }} + {{- with .Values.global.labels }} + {{- toYaml . | nindent 4 }} + {{- end }} + {{- if .Values.global.annotations }} + annotations: {{ toYaml .Values.global.annotations | nindent 4}} + {{- end }} +spec: + updateStrategy: + {{- with .Values.devicePlugin.updateStrategy }} + {{- toYaml . | nindent 4 }} + {{- end }} + selector: + matchLabels: + app.kubernetes.io/component: hami-device-plugin + {{- include "hami-vgpu.selectorLabels" . | nindent 6 }} + template: + metadata: + labels: + app.kubernetes.io/component: hami-device-plugin + hami.io/webhook: ignore + {{- include "hami-vgpu.selectorLabels" . | nindent 8 }} + annotations: + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + checksum/hami-scheduler-newversion-config: {{ include (print $.Template.BasePath "/scheduler/configmapnew.yaml") . | sha256sum }} + {{- else }} + checksum/hami-scheduler-config: {{ include (print $.Template.BasePath "/scheduler/configmap.yaml") . | sha256sum }} + {{- end }} + checksum/hami-scheduler-device-config: {{ include (print $.Template.BasePath "/scheduler/device-configmap.yaml") . | sha256sum }} + {{- if .Values.devicePlugin.podAnnotations }} + {{- toYaml .Values.devicePlugin.podAnnotations | nindent 8 }} + {{- end }} + spec: + {{- if .Values.devicePlugin.runtimeClassName }} + runtimeClassName: {{ .Values.devicePlugin.runtimeClassName }} + {{- end }} + serviceAccountName: {{ include "hami-vgpu.device-plugin" . }} + priorityClassName: system-node-critical + hostPID: true + hostNetwork: true + {{- include "hami.devicePlugin.imagePullSecrets" . | nindent 6 }} + {{- if .Values.devicePlugin.gpuOperatorToolkitReady.enabled }} + initContainers: + - name: toolkit-validation + image: {{ include "hami.devicePlugin.image" . }} + imagePullPolicy: {{ .Values.devicePlugin.image.pullPolicy }} + securityContext: + privileged: true + runAsUser: 0 + command: ["sh", "-c"] + args: + - | + echo "Waiting for NVIDIA Toolkit to be ready..." + until [ -f {{ .Values.devicePlugin.gpuOperatorToolkitReady.hostPath }}/toolkit-ready ]; do + echo "Waiting for {{ .Values.devicePlugin.gpuOperatorToolkitReady.hostPath }}/toolkit-ready..." + sleep 5 + done + echo "NVIDIA Toolkit is ready!" + volumeMounts: + - name: nvidia-validations + mountPath: {{ .Values.devicePlugin.gpuOperatorToolkitReady.hostPath | quote }} + mountPropagation: HostToContainer + readOnly: true + {{- end }} + containers: + - name: device-plugin + image: {{ include "hami.devicePlugin.image" . }} + imagePullPolicy: {{ .Values.devicePlugin.image.pullPolicy }} + lifecycle: + postStart: + exec: + command: ["/bin/sh","-c", {{ printf "/k8s-vgpu/bin/vgpu-init.sh %s/vgpu/" .Values.global.gpuHookPath | quote }}] + command: + - nvidia-device-plugin + - --config-file=/device-config.yaml + - --mig-strategy={{ .Values.devicePlugin.migStrategy }} + - --disable-core-limit={{ .Values.devicePlugin.disablecorelimit }} + {{- range .Values.devicePlugin.extraArgs }} + - {{ . }} + {{- end }} + env: + - name: NODE_NAME + valueFrom: + fieldRef: + fieldPath: spec.nodeName + - name: NVIDIA_MIG_MONITOR_DEVICES + value: all + - name: DEVICE_LIST_STRATEGY + value: {{ .Values.devicePlugin.deviceListStrategy }} + - name: HOOK_PATH + value: {{ .Values.global.gpuHookPath }} + {{- if typeIs "bool" .Values.devicePlugin.passDeviceSpecsEnabled }} + - name: PASS_DEVICE_SPECS + value: {{ .Values.devicePlugin.passDeviceSpecsEnabled | quote }} + {{- end }} + {{- if typeIs "string" .Values.devicePlugin.nvidiaDriverRoot }} + - name: NVIDIA_DRIVER_ROOT + value: {{ .Values.devicePlugin.nvidiaDriverRoot }} + {{- end }} + {{- if typeIs "string" .Values.devicePlugin.nvidiaHookPath }} + - name: NVIDIA_CDI_HOOK_PATH + value: {{ .Values.devicePlugin.nvidiaHookPath }} + {{- end }} + {{- if typeIs "bool" .Values.devicePlugin.gdrcopyEnabled }} + - name: GDRCOPY_ENABLED + value: {{ .Values.devicePlugin.gdrcopyEnabled | quote }} + {{- end }} + {{- if typeIs "bool" .Values.devicePlugin.gdsEnabled }} + - name: GDS_ENABLED + value: {{ .Values.devicePlugin.gdsEnabled | quote }} + {{- end }} + {{- if typeIs "bool" .Values.devicePlugin.mofedEnabled }} + - name: MOFED_ENABLED + value: {{ .Values.devicePlugin.mofedEnabled | quote }} + {{- end }} + {{- if eq (.Values.scheduler.defaultSchedulerPolicy.gpuSchedulerPolicy | default "spread") "topology-aware" }} + - name: ENABLE_TOPOLOGY_SCORE + value: "true" + {{- end }} + {{- with .Values.devicePlugin.extraEnvs }} + {{- . | toYaml | nindent 12 }} + {{- end }} + securityContext: + privileged: true + allowPrivilegeEscalation: true + capabilities: + drop: ["ALL"] + add: ["SYS_ADMIN"] + resources: + {{- toYaml .Values.devicePlugin.resources | nindent 12 }} + volumeMounts: + - name: device-plugin + mountPath: /var/lib/kubelet/device-plugins + - name: lib + mountPath: {{ printf "%s%s" .Values.global.gpuHookPath "/vgpu" }} + - name: usrbin + mountPath: /usrbin + - name: deviceconfig + mountPath: /config + - name: hosttmp + mountPath: /tmp + - name: device-config + mountPath: /device-config.yaml + subPath: device-config.yaml + - name: cdi-root + mountPath: /var/run/cdi + {{- if typeIs "string" .Values.devicePlugin.nvidiaDriverRoot }} + # We always mount the driver root at /driver-root in the container. + # This is required for CDI detection to work correctly. + - name: driver-root + mountPath: /driver-root + readOnly: true + {{- end }} + - name: vgpu-monitor + image: {{ include "hami.devicePlugin.monitor.image" . }} + imagePullPolicy: {{ .Values.devicePlugin.monitor.image.pullPolicy }} + command: + - "vGPUmonitor" + {{- range .Values.devicePlugin.monitor.extraArgs }} + - {{ . }} + {{- end }} + securityContext: + allowPrivilegeEscalation: false + capabilities: + drop: ["ALL"] + add: ["SYS_ADMIN"] + env: + - name: NODE_NAME + valueFrom: + fieldRef: + fieldPath: spec.nodeName + - name: NVIDIA_VISIBLE_DEVICES + value: "all" + - name: NVIDIA_MIG_MONITOR_DEVICES + value: "all" + - name: HOOK_PATH + value: "{{ .Values.global.gpuHookPath }}/vgpu" + - name: HAMI_RESYNC_INTERVAL + value: {{ .Values.devicePlugin.monitor.resyncInterval | default "5m" | quote }} + {{- with .Values.devicePlugin.monitor.extraEnvs }} + {{- . | toYaml | nindent 12 }} + {{- end }} + resources: + {{- toYaml .Values.devicePlugin.monitor.resources | nindent 12 }} + volumeMounts: + - name: ctrs + mountPath: {{ .Values.devicePlugin.monitor.ctrPath }} + - name: dockers + mountPath: /run/docker + - name: containerds + mountPath: /run/containerd + - name: sysinfo + mountPath: /sysinfo + - name: hostvar + mountPath: /hostvar + - name: hosttmp + mountPath: /tmp + volumes: + - name: ctrs + hostPath: + path: {{ .Values.devicePlugin.monitor.ctrPath }} + - name: hosttmp + hostPath: + path: /tmp + - name: dockers + hostPath: + path: /run/docker + - name: containerds + hostPath: + path: /run/containerd + - name: device-plugin + hostPath: + path: {{ .Values.devicePlugin.pluginPath }} + - name: lib + hostPath: + path: {{ .Values.devicePlugin.libPath }} + {{- if .Values.devicePlugin.gpuOperatorToolkitReady.enabled }} + - name: nvidia-validations + hostPath: + path: {{ .Values.devicePlugin.gpuOperatorToolkitReady.hostPath }} + type: DirectoryOrCreate + {{- end }} + {{- if typeIs "string" .Values.devicePlugin.nvidiaDriverRoot }} + - name: driver-root + hostPath: + path: {{ .Values.devicePlugin.nvidiaDriverRoot }} + type: Directory + {{- end }} + - name: cdi-root + hostPath: + path: /var/run/cdi + type: DirectoryOrCreate + - name: usrbin + hostPath: + path: /usr/bin + - name: sysinfo + hostPath: + path: /sys + - name: hostvar + hostPath: + path: /var + - name: deviceconfig + configMap: + name: {{ .Values.devicePlugin.nodeConfiguration.externalConfigName | default (include "hami-vgpu.device-plugin" .) }} + - name: device-config + configMap: + name: {{ include "hami-vgpu.scheduler" . }}-device + {{- if .Values.devicePlugin.nvidiaNodeSelector }} + nodeSelector: {{ toYaml .Values.devicePlugin.nvidiaNodeSelector | nindent 8 }} + {{- end }} + {{- if .Values.devicePlugin.tolerations }} + tolerations: {{ toYaml .Values.devicePlugin.tolerations | nindent 8 }} + {{- end }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/device-plugin/monitorrole.yaml b/packages/system/hami/charts/hami/templates/device-plugin/monitorrole.yaml new file mode 100644 index 00000000..6ac757f2 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/monitorrole.yaml @@ -0,0 +1,28 @@ +{{- if .Values.devicePlugin.enabled -}} +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRole +metadata: + name: {{ include "hami-vgpu.device-plugin" . }}-monitor +rules: + - apiGroups: + - "" + resources: + - pods + verbs: + - get + - create + - watch + - list + - update + - patch + - apiGroups: + - "" + resources: + - nodes + verbs: + - get + - update + - list + - patch +{{- end -}} + diff --git a/packages/system/hami/charts/hami/templates/device-plugin/monitorrolebinding.yaml b/packages/system/hami/charts/hami/templates/device-plugin/monitorrolebinding.yaml new file mode 100644 index 00000000..2f0a14ba --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/monitorrolebinding.yaml @@ -0,0 +1,17 @@ +{{- if .Values.devicePlugin.enabled -}} +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRoleBinding +metadata: + name: {{ include "hami-vgpu.device-plugin" . }} + labels: + app.kubernetes.io/component: "hami-device-plugin" + {{- include "hami-vgpu.labels" . | nindent 4 }} +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: ClusterRole + name: {{ include "hami-vgpu.device-plugin" . }}-monitor +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.device-plugin" . }} + namespace: {{ include "hami-vgpu.namespace" . }} +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/device-plugin/monitorservice.yaml b/packages/system/hami/charts/hami/templates/device-plugin/monitorservice.yaml new file mode 100644 index 00000000..88f1b214 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/monitorservice.yaml @@ -0,0 +1,29 @@ +{{- if .Values.devicePlugin.enabled -}} +apiVersion: v1 +kind: Service +metadata: + name: {{ include "hami-vgpu.device-plugin" . }}-monitor + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-device-plugin + {{- include "hami-vgpu.labels" . | nindent 4 }} + {{- with .Values.devicePlugin.service.labels }} + {{- toYaml . | nindent 4 }} + {{- end }} + {{- if .Values.devicePlugin.service.annotations }} # Use devicePlugin instead of scheduler + annotations: {{ toYaml .Values.devicePlugin.service.annotations | nindent 4 }} + {{- end }} +spec: + type: {{ .Values.devicePlugin.service.type | default "NodePort" }} # Default type is NodePort + ports: + - name: monitorport + port: {{ .Values.devicePlugin.service.httpPort | default 31992 }} # Default HTTP port is 31992 + targetPort: 9394 + {{- if eq (.Values.devicePlugin.service.type | default "NodePort") "NodePort" }} # If type is NodePort, set nodePort + nodePort: {{ .Values.devicePlugin.service.httpPort | default 31992 }} + {{- end }} + protocol: TCP + selector: + app.kubernetes.io/component: hami-device-plugin + {{- include "hami-vgpu.selectorLabels" . | nindent 4 }} +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/device-plugin/monitorserviceaccount.yaml b/packages/system/hami/charts/hami/templates/device-plugin/monitorserviceaccount.yaml new file mode 100644 index 00000000..6e3c2a44 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/monitorserviceaccount.yaml @@ -0,0 +1,10 @@ +{{- if .Values.devicePlugin.enabled -}} +apiVersion: v1 +kind: ServiceAccount +metadata: + name: {{ include "hami-vgpu.device-plugin" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: "hami-device-plugin" + {{- include "hami-vgpu.labels" . | nindent 4 }} +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/device-plugin/runtime-class.yaml b/packages/system/hami/charts/hami/templates/device-plugin/runtime-class.yaml new file mode 100644 index 00000000..ebcc2c9c --- /dev/null +++ b/packages/system/hami/charts/hami/templates/device-plugin/runtime-class.yaml @@ -0,0 +1,9 @@ +{{- if and .Values.devicePlugin.enabled .Values.devicePlugin.createRuntimeClass .Values.devicePlugin.runtimeClassName -}} +apiVersion: node.k8s.io/v1 +kind: RuntimeClass +metadata: + name: {{ .Values.devicePlugin.runtimeClassName }} + annotations: + helm.sh/hook: pre-install,pre-upgrade +handler: nvidia +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/scheduler/certmanager.yaml b/packages/system/hami/charts/hami/templates/scheduler/certmanager.yaml new file mode 100644 index 00000000..e6d28721 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/certmanager.yaml @@ -0,0 +1,31 @@ +{{- if .Values.scheduler.admissionWebhook.enabled -}} +{{- if .Values.scheduler.certManager.enabled }} +apiVersion: cert-manager.io/v1 +kind: Certificate +metadata: + name: {{ include "hami-vgpu.scheduler" . }}-serving-cert + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.labels" . | nindent 4 }} +spec: + dnsNames: + - {{ include "hami-vgpu.scheduler" . }}.{{ include "hami-vgpu.namespace" . }}.svc + - {{ include "hami-vgpu.scheduler" . }}.{{ include "hami-vgpu.namespace" . }}.svc.cluster.local + issuerRef: + kind: Issuer + name: {{ include "hami-vgpu.scheduler" . }}-selfsigned-issuer + secretName: {{ include "hami-vgpu.scheduler.tls" . }} +--- +apiVersion: cert-manager.io/v1 +kind: Issuer +metadata: + name: {{ include "hami-vgpu.scheduler" . }}-selfsigned-issuer + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.labels" . | nindent 4 }} +spec: + selfSigned: {} +{{- end }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/clusterrole.yaml b/packages/system/hami/charts/hami/templates/scheduler/clusterrole.yaml new file mode 100644 index 00000000..81c4fddf --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/clusterrole.yaml @@ -0,0 +1,34 @@ +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRole +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + labels: + app.kubernetes.io/component: "hami-scheduler" + {{- include "hami-vgpu.labels" . | nindent 4 }} +rules: + - apiGroups: [""] + resources: ["pods", "configmaps"] + verbs: ["get", "list", "watch", "patch"] + - apiGroups: [""] + resources: ["pods/binding"] + verbs: ["create"] + - apiGroups: [""] + resources: ["nodes"] + verbs: ["get", "list", "patch", "watch"] + - apiGroups: [""] + resources: ["events"] + verbs: ["create", "get", "list"] + - apiGroups: [""] + resources: ["resourcequotas"] + verbs: ["get", "list", "watch"] +--- +{{- if .Values.mockDevicePlugin.enabled }} +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRole +metadata: + name: {{ include "hami-vgpu.mock-device-plugin" . }} +rules: + - apiGroups: [""] + resources: ["nodes"] + verbs: ["get", "update", "list", "patch"] +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/scheduler/clusterrolebinding.yaml b/packages/system/hami/charts/hami/templates/scheduler/clusterrolebinding.yaml new file mode 100644 index 00000000..a81d425c --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/clusterrolebinding.yaml @@ -0,0 +1,47 @@ +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRoleBinding +metadata: + name: {{ include "hami-vgpu.scheduler" . }}-kube + labels: + app.kubernetes.io/component: "hami-scheduler" + {{- include "hami-vgpu.labels" . | nindent 4 }} +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: ClusterRole + name: system:kube-scheduler +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} +--- +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRoleBinding +metadata: + name: {{ include "hami-vgpu.scheduler" . }}-volume + labels: + app.kubernetes.io/component: "hami-scheduler" + {{- include "hami-vgpu.labels" . | nindent 4 }} +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: ClusterRole + name: system:volume-scheduler +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} +--- +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRoleBinding +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + labels: + app.kubernetes.io/component: "hami-scheduler" + {{- include "hami-vgpu.labels" . | nindent 4 }} +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: ClusterRole + name: {{ include "hami-vgpu.scheduler" . }} +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/configmap.yaml b/packages/system/hami/charts/hami/templates/scheduler/configmap.yaml new file mode 100644 index 00000000..109ecbce --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/configmap.yaml @@ -0,0 +1,142 @@ +{{- if .Values.scheduler.kubeScheduler.enabled -}} +apiVersion: v1 +kind: ConfigMap +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.labels" . | nindent 4 }} +data: + config.json: | + { + "kind": "Policy", + "apiVersion": "v1", + "extenders": [ + { + {{- if .Values.scheduler.admissionWebhook.enabled }} + "urlPrefix": "https://127.0.0.1:443", + "enableHttps": true, + "tlsConfig": { + "insecure": true + }, + {{- else }} + "urlPrefix": "http://127.0.0.1:80", + "enableHttps": false, + {{- end }} + "filterVerb": "filter", + "bindVerb": "bind", + "weight": 1, + "nodeCacheCapable": true, + "httpTimeout": 30000000000, + "managedResources": [ + {{- range .Values.devices.amd.customresources }} + { + "name": "{{ . }}", + "ignoredByScheduler": true + }, + {{- end }} + {{- if .Values.devices.ascend.enabled }} + {{- range .Values.devices.ascend.customresources }} + { + "name": "{{ . }}", + "ignoredByScheduler": true + }, + {{- end }} + {{- end }} + {{- if .Values.devices.mthreads.enabled }} + {{- range .Values.devices.mthreads.customresources }} + { + "name": "{{ . }}", + "ignoredByScheduler": true + }, + {{- end }} + {{- end }} + {{- if .Values.devices.enflame.enabled }} + {{- range .Values.devices.enflame.customresources }} + { + "name": "{{ . }}", + "ignoredByScheduler": true + }, + {{- end }} + {{- end }} + {{- if .Values.devices.kunlun.enabled }} + {{- range .Values.devices.kunlun.customresources }} + { + "name": "{{ . }}", + "ignoredByScheduler": true + }, + {{- end }} + {{- end }} + {{- range .Values.devices.awsneuron.customresources }} + { + "name": "{{ . }}", + "ignoredByScheduler": true + }, + {{- end }} + {{- if .Values.devices.iluvatar.enabled }} + {{- range .Values.devices.iluvatar.customresources }} + { + "name": "{{ . }}", + "ignoredByScheduler": true + }, + {{- end }} + {{- end }} + { + "name": "{{ .Values.resourceName }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.resourceMem }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.resourceCores }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.resourceMemPercentage }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.resourcePriority }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.mluResourceName }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.dcuResourceName }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.dcuResourceMem }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.dcuResourceCores }}", + "ignoredByScheduler": true + }, + { + "name": "metax-tech.com/gpu", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.metaxResourceName }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.metaxResourceCore }}", + "ignoredByScheduler": true + }, + { + "name": "{{ .Values.metaxResourceMem }}", + "ignoredByScheduler": true + } + ], + "ignoreable": false + } + ] + } +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/configmapnew.yaml b/packages/system/hami/charts/hami/templates/scheduler/configmapnew.yaml new file mode 100644 index 00000000..6f6db097 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/configmapnew.yaml @@ -0,0 +1,102 @@ +{{- if .Values.scheduler.kubeScheduler.enabled -}} +apiVersion: v1 +kind: ConfigMap +metadata: + name: {{ include "hami-vgpu.scheduler" . }}-newversion + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.labels" . | nindent 4 }} +data: + config.yaml: | + {{- if gt (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 25}} + apiVersion: kubescheduler.config.k8s.io/v1 + {{- else }} + apiVersion: kubescheduler.config.k8s.io/v1beta2 + {{- end }} + kind: KubeSchedulerConfiguration + leaderElection: + leaderElect: false + profiles: + - schedulerName: {{ .Values.schedulerName }} + extenders: + {{- if .Values.scheduler.admissionWebhook.enabled }} + - urlPrefix: "https://127.0.0.1:443" + enableHTTPS: true + tlsConfig: + insecure: true + {{- else }} + - urlPrefix: "http://127.0.0.1:80" + enableHTTPS: false + {{- end }} + filterVerb: filter + bindVerb: bind + nodeCacheCapable: true + weight: 1 + httpTimeout: 30s + managedResources: + - name: {{ .Values.resourceName }} + ignoredByScheduler: true + - name: {{ .Values.resourceMem }} + ignoredByScheduler: true + - name: {{ .Values.resourceCores }} + ignoredByScheduler: true + - name: {{ .Values.resourceMemPercentage }} + ignoredByScheduler: true + - name: {{ .Values.resourcePriority }} + ignoredByScheduler: true + - name: {{ .Values.mluResourceName }} + ignoredByScheduler: true + - name: {{ .Values.dcuResourceName }} + ignoredByScheduler: true + - name: {{ .Values.dcuResourceMem }} + ignoredByScheduler: true + - name: {{ .Values.dcuResourceCores }} + ignoredByScheduler: true + - name: "metax-tech.com/gpu" + ignoredByScheduler: true + - name: {{ .Values.metaxResourceName }} + ignoredByScheduler: true + - name: {{ .Values.metaxResourceCore }} + ignoredByScheduler: true + - name: {{ .Values.metaxResourceMem }} + ignoredByScheduler: true + {{- if .Values.devices.ascend.enabled }} + {{- range .Values.devices.ascend.customresources }} + - name: {{ . }} + ignoredByScheduler: true + {{- end }} + {{- end }} + {{- if .Values.devices.mthreads.enabled }} + {{- range .Values.devices.mthreads.customresources }} + - name: {{ . }} + ignoredByScheduler: true + {{- end }} + {{- end }} + {{- if .Values.devices.enflame.enabled }} + {{- range .Values.devices.enflame.customresources }} + - name: {{ . }} + ignoredByScheduler: true + {{- end }} + {{- end }} + {{- if .Values.devices.kunlun.enabled }} + {{- range .Values.devices.kunlun.customresources }} + - name: {{ . }} + ignoredByScheduler: true + {{- end }} + {{- end }} + {{- range .Values.devices.awsneuron.customresources }} + - name: {{ . }} + ignoredByScheduler: true + {{- end }} + {{- if .Values.devices.iluvatar.enabled }} + {{- range .Values.devices.iluvatar.customresources }} + - name: {{ . }} + ignoredByScheduler: true + {{- end }} + {{- end }} + {{- range .Values.devices.amd.customresources }} + - name: {{ . }} + ignoredByScheduler: true + {{- end }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/deployment.yaml b/packages/system/hami/charts/hami/templates/scheduler/deployment.yaml new file mode 100644 index 00000000..1d89e189 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/deployment.yaml @@ -0,0 +1,225 @@ +apiVersion: apps/v1 +kind: Deployment +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.labels" . | nindent 4 }} + {{- with .Values.global.labels }} + {{- toYaml . | nindent 4 }} + {{- end }} + {{- if .Values.global.annotations }} + annotations: {{ toYaml .Values.global.annotations | nindent 4}} + {{- end }} +spec: + {{- if .Values.scheduler.leaderElect }} + replicas: {{ .Values.scheduler.replicas }} + {{- else }} + replicas: 1 + {{- end }} + selector: + matchLabels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.selectorLabels" . | nindent 6 }} + template: + metadata: + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.selectorLabels" . | nindent 8 }} + hami.io/webhook: ignore + annotations: + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + checksum/hami-scheduler-newversion-config: {{ include (print $.Template.BasePath "/scheduler/configmapnew.yaml") . | sha256sum }} + {{- else }} + checksum/hami-scheduler-config: {{ include (print $.Template.BasePath "/scheduler/configmap.yaml") . | sha256sum }} + {{- end }} + checksum/hami-scheduler-device-config: {{ include (print $.Template.BasePath "/scheduler/device-configmap.yaml") . | sha256sum }} + {{- if .Values.scheduler.podAnnotations }} + {{- toYaml .Values.scheduler.podAnnotations | nindent 8 }} + {{- end }} + spec: + serviceAccountName: {{ include "hami-vgpu.scheduler" . }} + priorityClassName: system-node-critical + {{- include "hami.scheduler.extender.imagePullSecrets" . | nindent 6 }} + containers: + {{- if .Values.scheduler.kubeScheduler.enabled }} + - name: kube-scheduler + image: {{ include "hami.scheduler.kubeScheduler.image" . }} + imagePullPolicy: {{ .Values.scheduler.kubeScheduler.image.pullPolicy }} + command: + - kube-scheduler + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + {{- range .Values.scheduler.kubeScheduler.extraNewArgs }} + - {{ . }} + {{- end }} + {{- else }} + - --scheduler-name={{ .Values.schedulerName }} + {{- range .Values.scheduler.kubeScheduler.extraArgs }} + - {{ . }} + {{- end }} + {{- end }} + - --leader-elect={{ .Values.scheduler.leaderElect }} + - --leader-elect-resource-name={{ .Values.schedulerName }} + - --leader-elect-resource-namespace={{ include "hami-vgpu.namespace" . }} + resources: + {{- toYaml .Values.scheduler.kubeScheduler.resources | nindent 12 }} + volumeMounts: + - name: scheduler-config + mountPath: /config + {{- end }} + {{- if .Values.scheduler.livenessProbe }} + livenessProbe: + failureThreshold: 8 + httpGet: + path: /healthz + port: 10259 + scheme: HTTPS + initialDelaySeconds: 10 + periodSeconds: 10 + successThreshold: 1 + timeoutSeconds: 15 + {{- end }} + - name: vgpu-scheduler-extender + image: {{ include "hami.scheduler.extender.image" . }} + imagePullPolicy: {{ .Values.scheduler.extender.image.pullPolicy }} + env: + {{- if .Values.scheduler.nodeLockExpire }} + - name: HAMI_NODELOCK_EXPIRE + value: "{{ .Values.scheduler.nodeLockExpire }}" + {{- end }} + {{- if .Values.global.managedNodeSelectorEnable }} + {{- range $key, $value := .Values.global.managedNodeSelector }} + - name: NODE_SELECTOR_{{ $key | upper | replace "-" "_" }} + value: "{{ $value }}" + {{- end }} + {{- end }} + command: + - scheduler + {{- if .Values.scheduler.admissionWebhook.enabled }} + - --http_bind=0.0.0.0:443 + - --cert_file=/tls/tls.crt + - --key_file=/tls/tls.key + {{- else }} + - --http_bind=0.0.0.0:80 + {{- end }} + - --scheduler-name={{ .Values.schedulerName }} + - --metrics-bind-address={{ .Values.scheduler.metricsBindAddress }} + - --node-scheduler-policy={{ .Values.scheduler.defaultSchedulerPolicy.nodeSchedulerPolicy }} + - --gpu-scheduler-policy={{ .Values.scheduler.defaultSchedulerPolicy.gpuSchedulerPolicy }} + - --force-overwrite-default-scheduler={{ .Values.scheduler.forceOverwriteDefaultScheduler}} + - --device-config-file=/device-config.yaml + - --leader-elect={{ .Values.scheduler.leaderElect }} + - --leader-elect-resource-name={{ .Values.schedulerName }} + - --leader-elect-resource-namespace={{ include "hami-vgpu.namespace" . }} + {{- if .Values.devices.ascend.enabled }} + - --enable-ascend=true + {{- end }} + {{- if .Values.devices.iluvatar.enabled }} + - --enable-iluvatar=true + {{- end }} + {{- if .Values.scheduler.nodeLabelSelector }} + - --node-label-selector={{- $first := true -}} + {{- range $key, $value := .Values.scheduler.nodeLabelSelector -}} + {{- if not $first }},{{ end -}} + {{- $key }}={{ $value -}} + {{- $first = false -}} + {{- end -}} + {{- end }} + {{- range .Values.scheduler.extender.extraArgs }} + - {{ . }} + {{- end }} + ports: + {{- if .Values.scheduler.admissionWebhook.enabled }} + - name: https + containerPort: 443 + protocol: TCP + {{- else }} + - name: http + containerPort: 80 + protocol: TCP + {{- end }} + - name: metrics + containerPort: {{ last (splitList ":" .Values.scheduler.metricsBindAddress) | int }} + protocol: TCP + resources: + {{- toYaml .Values.scheduler.extender.resources | nindent 12 }} + volumeMounts: + - name: device-config + mountPath: /device-config.yaml + subPath: device-config.yaml + {{- if .Values.scheduler.admissionWebhook.enabled }} + - name: tls-config + mountPath: /tls + {{- end }} + {{- if .Values.scheduler.livenessProbe }} + livenessProbe: + httpGet: + path: /healthz + {{- if .Values.scheduler.admissionWebhook.enabled }} + port: https + scheme: HTTPS + {{- else }} + port: http + scheme: HTTP + {{- end }} + initialDelaySeconds: 10 + periodSeconds: 10 + failureThreshold: 3 + timeoutSeconds: 5 + {{- end }} + {{- if .Values.scheduler.leaderElect }} + readinessProbe: + httpGet: + path: /readyz + {{- if .Values.scheduler.admissionWebhook.enabled }} + port: https + scheme: HTTPS + {{- else }} + port: http + scheme: HTTP + {{- end }} + initialDelaySeconds: 10 + periodSeconds: 10 + failureThreshold: 3 + timeoutSeconds: 5 + {{- end }} + {{- if .Values.scheduler.leaderElect }} + affinity: + podAntiAffinity: + requiredDuringSchedulingIgnoredDuringExecution: + - labelSelector: + matchExpressions: + - key: app.kubernetes.io/component + operator: In + values: + - hami-scheduler + topologyKey: "kubernetes.io/hostname" + {{- end }} + volumes: + {{- if .Values.scheduler.admissionWebhook.enabled }} + - name: tls-config + secret: + secretName: {{ template "hami-vgpu.scheduler.tls" . }} + {{- end }} + {{- if .Values.scheduler.kubeScheduler.enabled }} + - name: scheduler-config + configMap: + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + name: {{ template "hami-vgpu.scheduler" . }}-newversion + {{- else }} + name: {{ template "hami-vgpu.scheduler" . }} + {{- end }} + {{- end }} + - name: device-config + configMap: + name: {{ include "hami-vgpu.scheduler" . }}-device + {{- if .Values.scheduler.nodeSelector }} + nodeSelector: {{ toYaml .Values.scheduler.nodeSelector | nindent 8 }} + {{- end }} + {{- if .Values.scheduler.tolerations }} + tolerations: {{ toYaml .Values.scheduler.tolerations | nindent 8 }} + {{- end }} + {{- if .Values.scheduler.nodeName }} + nodeName: {{ .Values.scheduler.nodeName }} + {{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/device-configmap.yaml b/packages/system/hami/charts/hami/templates/scheduler/device-configmap.yaml new file mode 100644 index 00000000..873b813b --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/device-configmap.yaml @@ -0,0 +1,408 @@ +apiVersion: v1 +kind: ConfigMap +metadata: + name: {{ include "hami-vgpu.scheduler" . }}-device + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.labels" . | nindent 4 }} +data: + device-config.yaml: |- + {{- if .Files.Glob "files/device-config.yaml" }} + {{- .Files.Get "files/device-config.yaml" | nindent 4}} + {{- else }} + nvidia: + resourceCountName: {{ .Values.resourceName }} + resourceMemoryName: {{ .Values.resourceMem }} + resourceMemoryPercentageName: {{ .Values.resourceMemPercentage }} + resourceCoreName: {{ .Values.resourceCores }} + resourcePriorityName: {{ .Values.resourcePriority }} + overwriteEnv: false + defaultMemory: 0 + defaultCores: 0 + defaultGPUNum: 1 + memoryFactor: 1 + deviceSplitCount: {{ .Values.devicePlugin.deviceSplitCount }} + deviceMemoryScaling: {{ .Values.devicePlugin.deviceMemoryScaling }} + deviceCoreScaling: {{ .Values.devicePlugin.deviceCoreScaling }} + gpuCorePolicy: {{ .Values.devices.nvidia.gpuCorePolicy }} + libCudaLogLevel: {{ .Values.devices.nvidia.libCudaLogLevel }} + runtimeClassName: "{{ .Values.devicePlugin.runtimeClassName }}" + knownMigGeometries: + - models: [ "A30" ] + allowedGeometries: + - + - name: 1g.6gb + core: 25 + memory: 6144 + count: 4 + - + - name: 2g.12gb + core: 50 + memory: 12288 + count: 2 + - + - name: 4g.24gb + core: 100 + memory: 24576 + count: 1 + - models: [ "A100-SXM4-40GB", "A100-40GB-PCIe", "A100-PCIE-40GB"] + allowedGeometries: + - + - name: 1g.5gb + core: 14 + memory: 5120 + count: 7 + - + - name: 1g.5gb + core: 14 + memory: 5120 + count: 1 + - name: 2g.10gb + core: 28 + memory: 10240 + count: 3 + - + - name: 3g.20gb + core: 42 + memory: 20480 + count: 2 + - + - name: 7g.40gb + core: 100 + memory: 40960 + count: 1 + - models: [ "A100-SXM4-80GB", "A100-80GB-PCIe", "A100-PCIE-80GB"] + allowedGeometries: + - + - name: 1g.10gb + core: 14 + memory: 10240 + count: 7 + - + - name: 1g.10gb + core: 14 + memory: 10240 + count: 1 + - name: 2g.20gb + core: 28 + memory: 20480 + count: 3 + - + - name: 3g.40gb + core: 42 + memory: 40960 + count: 2 + - + - name: 7g.79gb + core: 100 + memory: 80896 + count: 1 + - models: [ "H100-PCIE-80GB", "H100-SXM5-80GB"] + allowedGeometries: + - + - name: 1g.10gb + core: 14 + memory: 10240 + count: 7 + - + - name: 1g.10gb + core: 14 + memory: 10240 + count: 1 + - name: 2g.20gb + core: 28 + memory: 20480 + count: 3 + - + - name: 3g.40gb + core: 42 + memory: 40960 + count: 2 + - + - name: 7g.80gb + core: 100 + memory: 81920 + count: 1 + - models: [ "H100-PCIE-94GB", "H100-SXM5-94GB"] + allowedGeometries: + - + - name: 1g.12gb + core: 14 + memory: 12288 + count: 7 + - + - name: 1g.12gb + core: 14 + memory: 12288 + count: 1 + - name: 2g.24gb + core: 28 + memory: 24576 + count: 3 + - + - name: 3g.47gb + core: 42 + memory: 48128 + count: 2 + - + - name: 7g.94gb + core: 100 + memory: 96256 + count: 1 + - models: [ "H20", "H100 on GH200"] + allowedGeometries: + - + - name: 1g.12gb + core: 14 + memory: 12288 + count: 7 + - + - name: 1g.12gb + core: 14 + memory: 12288 + count: 1 + - name: 2g.24gb + core: 28 + memory: 24576 + count: 3 + - + - name: 3g.48gb + core: 42 + memory: 49152 + count: 2 + - + - name: 7g.96gb + core: 100 + memory: 98304 + count: 1 + - models: [ "H200 NVL", "H200-SXM5"] + allowedGeometries: + - + - name: 1g.18gb + core: 14 + memory: 18432 + count: 7 + - + - name: 1g.18gb + core: 14 + memory: 18432 + count: 1 + - name: 2g.35gb + core: 28 + memory: 35840 + count: 3 + - + - name: 3g.71gb + core: 42 + memory: 72704 + count: 2 + - + - name: 7g.141gb + core: 100 + memory: 144384 + count: 1 + - models: [ "B200" ] + allowedGeometries: + - + - name: 1g.23gb + core: 14 + memory: 23552 + count: 7 + - + - name: 1g.23gb + core: 14 + memory: 23552 + count: 1 + - name: 2g.45gb + core: 28 + memory: 46080 + count: 3 + - + - name: 3g.90gb + core: 42 + memory: 92160 + count: 2 + - + - name: 7g.180gb + core: 100 + memory: 184320 + count: 1 + cambricon: + resourceCountName: {{ .Values.mluResourceName }} + resourceMemoryName: {{ .Values.mluResourceMem }} + resourceCoreName: {{ .Values.mluResourceCores }} + hygon: + resourceCountName: {{ .Values.dcuResourceName }} + resourceMemoryName: {{ .Values.dcuResourceMem }} + resourceCoreName: {{ .Values.dcuResourceCores }} + memoryFactor: 1 + metax: + resourceCountName: "metax-tech.com/gpu" + resourceVCountName: {{ .Values.metaxResourceName }} + resourceVMemoryName: {{ .Values.metaxResourceMem }} + resourceVCoreName: {{ .Values.metaxResourceCore }} + sgpuTopologyAware: {{ .Values.metaxsGPUTopologyAware }} + enflame: + resourceNameGCU: "enflame.com/gcu" + resourceNameVGCU: {{ .Values.enflameResourceNameVGCU }} + resourceNameVGCUPercentage: {{ .Values.enflameResourceNameVGCUPercentage }} + mthreads: + resourceCountName: "mthreads.com/vgpu" + resourceMemoryName: "mthreads.com/sgpu-memory" + resourceCoreName: "mthreads.com/sgpu-core" + iluvatars: + - chipName: MR-V100 + commonWord: MR-V100 + resourceCountName: iluvatar.ai/MR-V100-vgpu + resourceMemoryName: iluvatar.ai/MR-V100.vMem + resourceCoreName: iluvatar.ai/MR-V100.vCore + - chipName: MR-V50 + commonWord: MR-V50 + resourceCountName: iluvatar.ai/MR-V50-vgpu + resourceMemoryName: iluvatar.ai/MR-V50.vMem + resourceCoreName: iluvatar.ai/MR-V50.vCore + - chipName: BI-V150 + commonWord: BI-V150 + resourceCountName: iluvatar.ai/BI-V150-vgpu + resourceMemoryName: iluvatar.ai/BI-V150.vMem + resourceCoreName: iluvatar.ai/BI-V150.vCore + - chipName: BI-V100 + commonWord: BI-V100 + resourceCountName: iluvatar.ai/BI-V100-vgpu + resourceMemoryName: iluvatar.ai/BI-V100.vMem + resourceCoreName: iluvatar.ai/BI-V100.vCore + kunlun: + resourceCountName: {{ .Values.kunlunResourceName }} + resourceVCountName: {{ .Values.kunlunResourceVCountName }} + resourceVMemoryName: {{ .Values.kunlunResourceVMemoryName }} + awsneuron: + resourceCountName: "aws.amazon.com/neuron" + resourceCoreName: "aws.amazon.com/neuroncore" + amd: + resourceCountName: "amd.com/gpu" + vnpus: + - chipName: 910A + commonWord: Ascend910A + resourceName: huawei.com/Ascend910A + resourceMemoryName: huawei.com/Ascend910A-memory + memoryAllocatable: 32768 + memoryCapacity: 32768 + memoryFactor: 1 + aiCore: 30 + templates: + - name: vir02 + memory: 2184 + aiCore: 2 + - name: vir04 + memory: 4369 + aiCore: 4 + - name: vir08 + memory: 8738 + aiCore: 8 + - name: vir16 + memory: 17476 + aiCore: 16 + - chipName: 910B2 + commonWord: Ascend910B2 + resourceName: huawei.com/Ascend910B2 + resourceMemoryName: huawei.com/Ascend910B2-memory + memoryAllocatable: 65536 + memoryCapacity: 65536 + memoryFactor: 1 + aiCore: 24 + aiCPU: 6 + templates: + - name: vir03_1c_8g + memory: 8192 + aiCore: 3 + aiCPU: 1 + - name: vir06_1c_16g + memory: 16384 + aiCore: 6 + aiCPU: 1 + - name: vir12_3c_32g + memory: 32768 + aiCore: 12 + aiCPU: 3 + - chipName: 910B3 + commonWord: Ascend910B3 + resourceName: huawei.com/Ascend910B3 + resourceMemoryName: huawei.com/Ascend910B3-memory + memoryAllocatable: 65536 + memoryCapacity: 65536 + memoryFactor: 1 + aiCore: 20 + aiCPU: 7 + templates: + - name: vir05_1c_16g + memory: 16384 + aiCore: 5 + aiCPU: 1 + - name: vir10_3c_32g + memory: 32768 + aiCore: 10 + aiCPU: 3 + - chipName: 910B4-1 + commonWord: Ascend910B4-1 + resourceName: huawei.com/Ascend910B4-1 + resourceMemoryName: huawei.com/Ascend910B4-1-memory + memoryAllocatable: 65536 + memoryCapacity: 65536 + memoryFactor: 1 + aiCore: 20 + aiCPU: 7 + templates: + # NOTE: Names of vnpu template for 910B4-1 are fixed by Ascend runtime and must not be changed. + # The memory is used for scheduling so the correct values must be set. + # Template vir05_1c_8g actually provides 16GB memory, + - name: vir05_1c_8g + memory: 16384 + aiCore: 5 + aiCPU: 1 + # Template vir10_3c_16g actually provides 32GB memory + - name: vir10_3c_16g + memory: 32768 + aiCore: 10 + aiCPU: 3 + - chipName: 910B4 + commonWord: Ascend910B4 + resourceName: huawei.com/Ascend910B4 + resourceMemoryName: huawei.com/Ascend910B4-memory + memoryAllocatable: 32768 + memoryCapacity: 32768 + memoryFactor: 1 + aiCore: 20 + aiCPU: 7 + templates: + - name: vir05_1c_8g + memory: 8192 + aiCore: 5 + aiCPU: 1 + - name: vir10_3c_16g + memory: 16384 + aiCore: 10 + aiCPU: 3 + - chipName: 310P3 + commonWord: Ascend310P + resourceName: huawei.com/Ascend310P + resourceMemoryName: huawei.com/Ascend310P-memory + memoryAllocatable: 21527 + memoryCapacity: 24576 + memoryFactor: 1 + aiCore: 8 + aiCPU: 7 + templates: + - name: vir01 + memory: 3072 + aiCore: 1 + aiCPU: 1 + - name: vir02 + memory: 6144 + aiCore: 2 + aiCPU: 2 + - name: vir04 + memory: 12288 + aiCore: 4 + aiCPU: 4 + {{ end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/job-patch/clusterrole.yaml b/packages/system/hami/charts/hami/templates/scheduler/job-patch/clusterrole.yaml new file mode 100644 index 00000000..77e891cc --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/job-patch/clusterrole.yaml @@ -0,0 +1,21 @@ +{{- if and .Values.scheduler.admissionWebhook.enabled (.Values.scheduler.patch.enabled) (not .Values.scheduler.certManager.enabled) -}} +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRole +metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission + annotations: + "helm.sh/hook": pre-install,pre-upgrade,post-install,post-upgrade + "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded + labels: + {{- include "hami-vgpu.labels" . | nindent 4 }} + app.kubernetes.io/component: admission-webhook +rules: + - apiGroups: + - admissionregistration.k8s.io + resources: + #- validatingwebhookconfigurations + - mutatingwebhookconfigurations + verbs: + - get + - update +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/job-patch/clusterrolebinding.yaml b/packages/system/hami/charts/hami/templates/scheduler/job-patch/clusterrolebinding.yaml new file mode 100644 index 00000000..2b82f926 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/job-patch/clusterrolebinding.yaml @@ -0,0 +1,20 @@ +{{- if and .Values.scheduler.admissionWebhook.enabled (.Values.scheduler.patch.enabled) (not .Values.scheduler.certManager.enabled) -}} +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRoleBinding +metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission + annotations: + "helm.sh/hook": pre-install,pre-upgrade,post-install,post-upgrade + "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded + labels: + {{- include "hami-vgpu.labels" . | nindent 4 }} + app.kubernetes.io/component: admission-webhook +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: ClusterRole + name: {{ include "hami-vgpu.fullname" . }}-admission +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.fullname" . }}-admission + namespace: {{ include "hami-vgpu.namespace" . }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/job-patch/job-createSecret.yaml b/packages/system/hami/charts/hami/templates/scheduler/job-patch/job-createSecret.yaml new file mode 100644 index 00000000..0e36d95f --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/job-patch/job-createSecret.yaml @@ -0,0 +1,68 @@ +{{- if and .Values.scheduler.admissionWebhook.enabled (.Values.scheduler.patch.enabled) (not .Values.scheduler.certManager.enabled) -}} +apiVersion: batch/v1 +kind: Job +metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission-create + namespace: {{ include "hami-vgpu.namespace" . }} + annotations: + "helm.sh/hook": pre-install,pre-upgrade + "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded + labels: + {{- include "hami-vgpu.labels" . | nindent 4 }} + app.kubernetes.io/component: admission-webhook +spec: + {{- if .Capabilities.APIVersions.Has "batch/v1alpha1" }} + # Alpha feature since k8s 1.12 + ttlSecondsAfterFinished: 0 + {{- end }} + template: + metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission-create + {{- if .Values.scheduler.patch.podAnnotations }} + annotations: {{ toYaml .Values.scheduler.patch.podAnnotations | nindent 8 }} + {{- end }} + labels: + {{- include "hami-vgpu.labels" . | nindent 8 }} + app.kubernetes.io/component: admission-webhook + hami.io/webhook: ignore + spec: + {{- if .Values.scheduler.patch.priorityClassName }} + priorityClassName: {{ .Values.scheduler.patch.priorityClassName }} + {{- end }} + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + {{- include "hami.scheduler.patch.new.imagePullSecrets" . | nindent 6 }} + {{- else }} + {{- include "hami.scheduler.patch.imagePullSecrets" . | nindent 6 }} + {{- end }} + containers: + - name: create + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + image: {{ include "hami.scheduler.patch.new.image" . }} + imagePullPolicy: {{ .Values.scheduler.patch.imageNew.pullPolicy }} + {{- else }} + image: {{ include "hami.scheduler.patch.image" . }} + imagePullPolicy: {{ .Values.scheduler.patch.image.pullPolicy }} + {{- end }} + args: + - create + - --cert-name=tls.crt + - --key-name=tls.key + {{- if .Values.scheduler.admissionWebhook.customURL.enabled }} + - --host={{ printf "%s.%s.svc,127.0.0.1,%s" (include "hami-vgpu.scheduler" .) (include "hami-vgpu.namespace" .) .Values.scheduler.admissionWebhook.customURL.host}} + {{- else }} + - --host={{ printf "%s.%s.svc,127.0.0.1" (include "hami-vgpu.scheduler" .) (include "hami-vgpu.namespace" .) }} + {{- end }} + - --namespace={{ include "hami-vgpu.namespace" . }} + - --secret-name={{ include "hami-vgpu.scheduler.tls" . }} + restartPolicy: OnFailure + serviceAccountName: {{ include "hami-vgpu.fullname" . }}-admission + {{- if .Values.scheduler.patch.nodeSelector }} + nodeSelector: {{ toYaml .Values.scheduler.patch.nodeSelector | nindent 8 }} + {{- end }} + {{- if .Values.scheduler.patch.tolerations }} + tolerations: {{ toYaml .Values.scheduler.patch.tolerations | nindent 8 }} + {{- end }} + securityContext: + runAsNonRoot: true + runAsUser: {{ .Values.scheduler.patch.runAsUser }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/job-patch/job-patchWebhook.yaml b/packages/system/hami/charts/hami/templates/scheduler/job-patch/job-patchWebhook.yaml new file mode 100644 index 00000000..ce52042c --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/job-patch/job-patchWebhook.yaml @@ -0,0 +1,63 @@ +{{- if and .Values.scheduler.admissionWebhook.enabled (.Values.scheduler.patch.enabled) (not .Values.scheduler.certManager.enabled) -}} +apiVersion: batch/v1 +kind: Job +metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission-patch + namespace: {{ include "hami-vgpu.namespace" . }} + annotations: + "helm.sh/hook": post-install,post-upgrade + "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded + labels: + {{- include "hami-vgpu.labels" . | nindent 4 }} + app.kubernetes.io/component: admission-webhook +spec: + {{- if .Capabilities.APIVersions.Has "batch/v1alpha1" }} + # Alpha feature since k8s 1.12 + ttlSecondsAfterFinished: 0 + {{- end }} + template: + metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission-patch + {{- if .Values.scheduler.patch.podAnnotations }} + annotations: {{ toYaml .Values.scheduler.patch.podAnnotations | nindent 8 }} + {{- end }} + labels: + {{- include "hami-vgpu.labels" . | nindent 8 }} + app.kubernetes.io/component: admission-webhook + hami.io/webhook: ignore + spec: + {{- if .Values.scheduler.patch.priorityClassName }} + priorityClassName: {{ .Values.scheduler.patch.priorityClassName }} + {{- end }} + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + {{- include "hami.scheduler.patch.new.imagePullSecrets" . | nindent 6 }} + {{- else }} + {{- include "hami.scheduler.patch.imagePullSecrets" . | nindent 6 }} + {{- end }} + containers: + - name: patch + {{- if ge (regexReplaceAll "[^0-9]" .Capabilities.KubeVersion.Minor "" | int) 22 }} + image: {{ include "hami.scheduler.patch.new.image" . }} + imagePullPolicy: {{ .Values.scheduler.patch.imageNew.pullPolicy }} + {{- else }} + image: {{ include "hami.scheduler.patch.image" . }} + imagePullPolicy: {{ .Values.scheduler.patch.image.pullPolicy }} + {{- end }} + args: + - patch + - --webhook-name={{ include "hami-vgpu.scheduler.webhook" . }} + - --namespace={{ include "hami-vgpu.namespace" . }} + - --patch-validating=false + - --secret-name={{ include "hami-vgpu.scheduler.tls" . }} + restartPolicy: OnFailure + serviceAccountName: {{ include "hami-vgpu.fullname" . }}-admission + {{- if .Values.scheduler.patch.nodeSelector }} + nodeSelector: {{ toYaml .Values.scheduler.patch.nodeSelector | nindent 8 }} + {{- end }} + {{- if .Values.scheduler.patch.tolerations }} + tolerations: {{ toYaml .Values.scheduler.patch.tolerations | nindent 8 }} + {{- end }} + securityContext: + runAsNonRoot: true + runAsUser: {{ .Values.scheduler.patch.runAsUser }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/job-patch/role.yaml b/packages/system/hami/charts/hami/templates/scheduler/job-patch/role.yaml new file mode 100644 index 00000000..56682f8e --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/job-patch/role.yaml @@ -0,0 +1,21 @@ +{{- if and .Values.scheduler.admissionWebhook.enabled (.Values.scheduler.patch.enabled) (not .Values.scheduler.certManager.enabled) -}} +apiVersion: rbac.authorization.k8s.io/v1 +kind: Role +metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission + namespace: {{ include "hami-vgpu.namespace" . }} + annotations: + "helm.sh/hook": pre-install,pre-upgrade,post-install,post-upgrade + "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded + labels: + {{- include "hami-vgpu.labels" . | nindent 4 }} + app.kubernetes.io/component: admission-webhook +rules: + - apiGroups: + - "" + resources: + - secrets + verbs: + - get + - create +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/job-patch/rolebinding.yaml b/packages/system/hami/charts/hami/templates/scheduler/job-patch/rolebinding.yaml new file mode 100644 index 00000000..7239b128 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/job-patch/rolebinding.yaml @@ -0,0 +1,21 @@ +{{- if and .Values.scheduler.admissionWebhook.enabled (.Values.scheduler.patch.enabled) (not .Values.scheduler.certManager.enabled) -}} +apiVersion: rbac.authorization.k8s.io/v1 +kind: RoleBinding +metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission + namespace: {{ include "hami-vgpu.namespace" . }} + annotations: + "helm.sh/hook": pre-install,pre-upgrade,post-install,post-upgrade + "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded + labels: + {{- include "hami-vgpu.labels" . | nindent 4 }} + app.kubernetes.io/component: admission-webhook +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: Role + name: {{ include "hami-vgpu.fullname" . }}-admission +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.fullname" . }}-admission + namespace: {{ include "hami-vgpu.namespace" . }} +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/job-patch/serviceaccount.yaml b/packages/system/hami/charts/hami/templates/scheduler/job-patch/serviceaccount.yaml new file mode 100644 index 00000000..857c6a8d --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/job-patch/serviceaccount.yaml @@ -0,0 +1,13 @@ +{{- if and .Values.scheduler.admissionWebhook.enabled (.Values.scheduler.patch.enabled) (not .Values.scheduler.certManager.enabled) }} +apiVersion: v1 +kind: ServiceAccount +metadata: + name: {{ include "hami-vgpu.fullname" . }}-admission + namespace: {{ include "hami-vgpu.namespace" . }} + annotations: + "helm.sh/hook": pre-install,pre-upgrade,post-install,post-upgrade + "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded + labels: + {{- include "hami-vgpu.labels" . | nindent 4 }} + app.kubernetes.io/component: admission-webhook +{{- end }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/role.yaml b/packages/system/hami/charts/hami/templates/scheduler/role.yaml new file mode 100644 index 00000000..5f7d7e44 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/role.yaml @@ -0,0 +1,12 @@ +apiVersion: rbac.authorization.k8s.io/v1 +kind: Role +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: "hami-scheduler" + {{- include "hami-vgpu.labels" . | nindent 4 }} +rules: + - apiGroups: ["coordination.k8s.io"] + resources: ["leases"] + verbs: ["create", "list", "watch", "get", "update", "patch"] diff --git a/packages/system/hami/charts/hami/templates/scheduler/rolebinding.yaml b/packages/system/hami/charts/hami/templates/scheduler/rolebinding.yaml new file mode 100644 index 00000000..96e175a1 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/rolebinding.yaml @@ -0,0 +1,31 @@ +apiVersion: rbac.authorization.k8s.io/v1 +kind: RoleBinding +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: "hami-scheduler" + {{- include "hami-vgpu.labels" . | nindent 4 }} +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: Role + name: {{ include "hami-vgpu.scheduler" . }} +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} +--- +{{- if .Values.mockDevicePlugin.enabled }} +apiVersion: rbac.authorization.k8s.io/v1 +kind: ClusterRoleBinding +metadata: + name: {{ include "hami-vgpu.mock-device-plugin" . }} +roleRef: + apiGroup: rbac.authorization.k8s.io + kind: ClusterRole + name: {{ include "hami-vgpu.mock-device-plugin" . }} +subjects: + - kind: ServiceAccount + name: {{ include "hami-vgpu.mock-device-plugin" . }} + namespace: {{ include "hami-vgpu.namespace" . }} +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/scheduler/service.yaml b/packages/system/hami/charts/hami/templates/scheduler/service.yaml new file mode 100644 index 00000000..d7538fed --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/service.yaml @@ -0,0 +1,34 @@ +apiVersion: v1 +kind: Service +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.labels" . | nindent 4 }} + {{- if .Values.scheduler.service.labels }} + {{ toYaml .Values.scheduler.service.labels | indent 4 }} + {{- end }} + {{- if .Values.scheduler.service.annotations }} + annotations: {{ toYaml .Values.scheduler.service.annotations | nindent 4 }} + {{- end }} +spec: + type: {{ .Values.scheduler.service.type | default "NodePort" }} # Default type is NodePort + ports: + - name: http + port: {{ .Values.scheduler.service.httpPort | default 443 }} # Default HTTP port is 443 + targetPort: {{ .Values.scheduler.service.httpTargetPort | default 443 }} + {{- if eq (.Values.scheduler.service.type | default "NodePort") "NodePort" }} # If type is NodePort, set nodePort + nodePort: {{ .Values.scheduler.service.schedulerPort | default 31998 }} + {{- end }} + protocol: TCP + - name: monitor + port: {{ .Values.scheduler.service.monitorPort | default 31993 }} # Default monitoring port is 31993 + targetPort: {{ .Values.scheduler.service.monitorTargetPort | default 9395 }} + {{- if eq (.Values.scheduler.service.type | default "NodePort") "NodePort" }} # If type is NodePort, set nodePort + nodePort: {{ .Values.scheduler.service.monitorPort | default 31993 }} + {{- end }} + protocol: TCP + selector: + app.kubernetes.io/component: hami-scheduler + {{- include "hami-vgpu.selectorLabels" . | nindent 4 }} diff --git a/packages/system/hami/charts/hami/templates/scheduler/serviceaccount.yaml b/packages/system/hami/charts/hami/templates/scheduler/serviceaccount.yaml new file mode 100644 index 00000000..0435c003 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/serviceaccount.yaml @@ -0,0 +1,16 @@ +apiVersion: v1 +kind: ServiceAccount +metadata: + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + labels: + app.kubernetes.io/component: "hami-scheduler" + {{- include "hami-vgpu.labels" . | nindent 4 }} +--- +{{- if .Values.mockDevicePlugin.enabled }} +apiVersion: v1 +kind: ServiceAccount +metadata: + name: {{ include "hami-vgpu.mock-device-plugin" . }} + namespace: {{ include "hami-vgpu.namespace" . }} +{{- end -}} diff --git a/packages/system/hami/charts/hami/templates/scheduler/webhook.yaml b/packages/system/hami/charts/hami/templates/scheduler/webhook.yaml new file mode 100644 index 00000000..db9f8029 --- /dev/null +++ b/packages/system/hami/charts/hami/templates/scheduler/webhook.yaml @@ -0,0 +1,57 @@ +{{- if .Values.scheduler.admissionWebhook.enabled -}} +apiVersion: admissionregistration.k8s.io/v1 +kind: MutatingWebhookConfiguration +metadata: + {{- if .Values.scheduler.certManager.enabled }} + annotations: + cert-manager.io/inject-ca-from: {{ include "hami-vgpu.namespace" . }}/{{ include "hami-vgpu.scheduler" . }}-serving-cert + {{- end }} + name: {{ include "hami-vgpu.scheduler.webhook" . }} +webhooks: + - admissionReviewVersions: + - v1beta1 + clientConfig: + {{- if .Values.scheduler.admissionWebhook.customURL.enabled }} + url: https://{{ .Values.scheduler.admissionWebhook.customURL.host}}:{{.Values.scheduler.admissionWebhook.customURL.port}}{{.Values.scheduler.admissionWebhook.customURL.path}} + {{- else }} + service: + name: {{ include "hami-vgpu.scheduler" . }} + namespace: {{ include "hami-vgpu.namespace" . }} + path: /webhook + port: {{ .Values.scheduler.service.httpPort }} + {{- end }} + failurePolicy: {{ .Values.scheduler.admissionWebhook.failurePolicy }} + matchPolicy: Equivalent + name: vgpu.hami.io + namespaceSelector: + matchExpressions: + - key: hami.io/webhook + operator: NotIn + values: + - ignore + {{- if .Values.scheduler.admissionWebhook.whitelistNamespaces }} + - key: kubernetes.io/metadata.name + operator: NotIn + values: + {{- toYaml .Values.scheduler.admissionWebhook.whitelistNamespaces | nindent 10 }} + {{- end }} + objectSelector: + matchExpressions: + - key: hami.io/webhook + operator: NotIn + values: + - ignore + reinvocationPolicy: {{ .Values.scheduler.admissionWebhook.reinvocationPolicy }} + rules: + - apiGroups: + - "" + apiVersions: + - v1 + operations: + - CREATE + resources: + - pods + scope: '*' + sideEffects: None + timeoutSeconds: 10 +{{- end }} diff --git a/packages/system/hami/charts/hami/values.yaml b/packages/system/hami/charts/hami/values.yaml new file mode 100644 index 00000000..01caca0b --- /dev/null +++ b/packages/system/hami/charts/hami/values.yaml @@ -0,0 +1,455 @@ +## @section Global configuration +## @param global.imageRegistry Global Docker image registry +## @param global.imagePullSecrets Global Docker image pull secrets +global: + ## @param global.imageRegistry Global Docker image registry + imageRegistry: "" + ## E.g. + ## imagePullSecrets: + ## - myRegistryKeySecretName + ## @param global.imagePullSecrets Global Docker image pull secrets + imagePullSecrets: [] + imageTag: "v2.8.1" + gpuHookPath: /usr/local + labels: {} + annotations: {} + managedNodeSelectorEnable: false + managedNodeSelector: + usage: "gpu" + +nameOverride: "" +fullnameOverride: "" +namespaceOverride: "" + +#Nvidia GPU Parameters +resourceName: "nvidia.com/gpu" +resourceMem: "nvidia.com/gpumem" +resourceMemPercentage: "nvidia.com/gpumem-percentage" +resourceCores: "nvidia.com/gpucores" +resourcePriority: "nvidia.com/priority" + +#MLU Parameters +mluResourceName: "cambricon.com/vmlu" +mluResourceMem: "cambricon.com/mlu.smlu.vmemory" +mluResourceCores: "cambricon.com/mlu.smlu.vcore" + +#Hygon DCU Parameters +dcuResourceName: "hygon.com/dcunum" +dcuResourceMem: "hygon.com/dcumem" +dcuResourceCores: "hygon.com/dcucores" + +#Metax sGPU Parameters +metaxResourceName: "metax-tech.com/sgpu" +metaxResourceCore: "metax-tech.com/vcore" +metaxResourceMem: "metax-tech.com/vmemory" +metaxsGPUTopologyAware: "false" + +#Enflame VGCU Parameters +enflameResourceNameVGCU: "enflame.com/vgcu" +enflameResourceNameVGCUPercentage: "enflame.com/vgcu-percentage" + +#Kunlun XPU Parameters +kunlunResourceName: "kunlunxin.com/xpu" +kunlunResourceVCountName: "kunlunxin.com/vxpu" +kunlunResourceVMemoryName: "kunlunxin.com/vxpu-memory" + +schedulerName: "hami-scheduler" + +scheduler: + # @param nodeName defines the node name and the nvidia-vgpu-scheduler-scheduler will schedule to the node. + # if we install the nvidia-vgpu-scheduler-scheduler as default scheduler, we need to remove the k8s default + # scheduler pod from the cluster first, we must specify node name to skip the schedule workflow. + nodeName: "" + #nodeLabelSelector: + # "gpu": "on" + overwriteEnv: "false" + defaultSchedulerPolicy: + nodeSchedulerPolicy: binpack + gpuSchedulerPolicy: spread + metricsBindAddress: ":9395" + # If set to false, When Pod.Spec.SchedulerName equals to the const DefaultSchedulerName in k8s.io/api/core/v1 package, webhook will not overwrite it + forceOverwriteDefaultScheduler: true + livenessProbe: false + leaderElect: true + # when leaderElect is true, replicas is available, otherwise replicas is 1. + replicas: 1 + kubeScheduler: + # @param enabled indicate whether to run kube-scheduler container in the scheduler pod, it's true by default. + enabled: true + ## @param image.registry kube scheduler image registry + ## @param image.repository kube scheduler image repository + ## @param image.tag kube scheduler image tag (immutable tags are recommended) + ## @param image.pullPolicy kube scheduler image pull policy + ## @param image.pullSecrets Specify docker-registry secret names as an array + image: + registry: "registry.cn-hangzhou.aliyuncs.com" + repository: "google_containers/kube-scheduler" + tag: "" + ## Specify a imagePullPolicy + ## Defaults to 'Always' if image tag is 'latest', else set to 'IfNotPresent' + ## ref: https://kubernetes.io/docs/user-guide/images/#pre-pulling-images + ## + pullPolicy: IfNotPresent + ## Optionally specify an array of imagePullSecrets. + ## Secrets must be manually created in the namespace. + ## Example: + ## pullSecrets: + ## - myRegistryKeySecretName + ## + pullSecrets: [] + resources: {} + # If you do want to specify resources, uncomment the following lines, adjust them as necessary. + # and remove the curly braces after 'resources:'. +# limits: +# cpu: 1000m +# memory: 1000Mi +# requests: +# cpu: 100m +# memory: 100Mi + extraNewArgs: + - --config=/config/config.yaml + - -v=4 + extraArgs: + - --policy-config-file=/config/config.json + - -v=4 + extender: + ## @param image.registry scheduler extender image registry + ## @param image.repository scheduler extender image repository + ## @param image.tag scheduler extender image tag (immutable tags are recommended) + ## @param image.pullPolicy scheduler extender image pull policy + ## @param image.pullSecrets Specify docker-registry secret names as an array + image: + registry: "docker.io" + repository: "projecthami/hami" + tag: "" + ## Specify a imagePullPolicy + ## Defaults to 'Always' if image tag is 'latest', else set to 'IfNotPresent' + ## ref: https://kubernetes.io/docs/user-guide/images/#pre-pulling-images + ## + pullPolicy: IfNotPresent + ## Optionally specify an array of imagePullSecrets. + ## Secrets must be manually created in the namespace. + ## Example: + ## pullSecrets: + ## - myRegistryKeySecretName + ## + pullSecrets: [] + resources: {} + # If you do want to specify resources, uncomment the following lines, adjust them as necessary, + # and remove the curly braces after 'resources:'. +# limits: +# cpu: 1000m +# memory: 1000Mi +# requests: +# cpu: 100m +# memory: 100Mi + extraArgs: + - --debug + - -v=4 + nodeLockExpire: "5m" + podAnnotations: {} + tolerations: [] + #serviceAccountName: "hami-vgpu-scheduler-sa" + admissionWebhook: + # If set to false, the admission webhook is not installed and any pods that should use HAMi must be + # configured to use the right 'schedulerName' and other device-specific configurations. + enabled: true + customURL: + enabled: false + # must be an endpoint using https. + # should generate host certs here + host: 127.0.0.1 # hostname or ip, can be your node'IP if you want to use https://:/ + port: 31998 + path: /webhook + whitelistNamespaces: + # Specify the namespaces that the webhook will not be applied to. + # - default + # - kube-system + # - istio-system + reinvocationPolicy: Never + failurePolicy: Ignore + ## TLS Certificate Option 1: Use cert-manager to generate self-signed certificate. + ## If enabled, always takes precedence over options 2. + certManager: + enabled: false + ## TLS Certificate Option 2: Use kube-webhook-certgen to generate self-signed certificate. + ## If true and certManager.enabled is false, Helm will automatically create a self-signed cert and secret for you. + patch: + enabled: true + ## @param image.registry scheduler extender image registry + ## @param image.repository scheduler extender image repository + ## @param image.tag scheduler extender image tag (immutable tags are recommended) + ## @param image.pullPolicy scheduler extender image pull policy + ## @param image.pullSecrets Specify docker-registry secret names as an array + image: + registry: "docker.io" + repository: "jettech/kube-webhook-certgen" + tag: "v1.5.2" + ## Specify a imagePullPolicy + ## Defaults to 'Always' if image tag is 'latest', else set to 'IfNotPresent' + ## ref: https://kubernetes.io/docs/user-guide/images/#pre-pulling-images + ## + pullPolicy: IfNotPresent + ## Optionally specify an array of imagePullSecrets. + ## Secrets must be manually created in the namespace. + ## Example: + ## pullSecrets: + ## - myRegistryKeySecretName + ## + pullSecrets: [] + ## @param image.registry scheduler extender image registry + ## @param image.repository scheduler extender image repository + ## @param image.tag scheduler extender image tag (immutable tags are recommended) + ## @param image.pullPolicy scheduler extender image pull policy + ## @param image.pullSecrets Specify docker-registry secret names as an array + imageNew: + registry: "docker.io" + repository: "liangjw/kube-webhook-certgen" + tag: "v1.1.1" + ## Specify a imagePullPolicy + ## Defaults to 'Always' if image tag is 'latest', else set to 'IfNotPresent' + ## ref: https://kubernetes.io/docs/user-guide/images/#pre-pulling-images + ## + pullPolicy: IfNotPresent + ## Optionally specify an array of imagePullSecrets. + ## Secrets must be manually created in the namespace. + ## Example: + ## pullSecrets: + ## - myRegistryKeySecretName + ## + pullSecrets: [] + priorityClassName: "" + podAnnotations: {} + nodeSelector: {} + tolerations: [] + runAsUser: 2000 + service: + type: NodePort # Default type is NodePort, can be changed to ClusterIP + httpPort: 443 # HTTP port + schedulerPort: 31998 # NodePort for HTTP + monitorPort: 31993 # Monitoring port + monitorTargetPort: 9395 + httpTargetPort: 443 + labels: {} + annotations: {} + +devicePlugin: + enabled: true + gpuOperatorToolkitReady: + enabled: false + hostPath: "/run/nvidia/validations" + ## @param image.registry devicePlugin image registry + ## @param image.repository devicePlugin image repository + ## @param image.tag devicePlugin image tag (immutable tags are recommended) + ## @param image.pullPolicy devicePlugin image pull policy + ## @param image.pullSecrets Specify docker-registry secret names as an array + image: + registry: "docker.io" + repository: "projecthami/hami" + tag: "" + ## Specify a imagePullPolicy + ## Defaults to 'Always' if image tag is 'latest', else set to 'IfNotPresent' + ## ref: https://kubernetes.io/docs/user-guide/images/#pre-pulling-images + ## + pullPolicy: IfNotPresent + ## Optionally specify an array of imagePullSecrets. + ## Secrets must be manually created in the namespace. + ## Example: + ## pullSecrets: + ## - myRegistryKeySecretName + ## + pullSecrets: [] + monitor: + ## @param image.registry monitor image registry + ## @param image.repository monitor image repository + ## @param image.tag monitor image tag (immutable tags are recommended) + ## @param image.pullPolicy monitor image pull policy + ## @param image.pullSecrets Specify docker-registry secret names as an array + image: + registry: "docker.io" + repository: "projecthami/hami" + tag: "" + ## Specify a imagePullPolicy + ## Defaults to 'Always' if image tag is 'latest', else set to 'IfNotPresent' + ## ref: https://kubernetes.io/docs/user-guide/images/#pre-pulling-images + ## + pullPolicy: IfNotPresent + ## Optionally specify an array of imagePullSecrets. + ## Secrets must be manually created in the namespace. + ## Example: + ## pullSecrets: + ## - myRegistryKeySecretName + ## + pullSecrets: [] + ctrPath: /usr/local/vgpu/containers + resyncInterval: "5m" + extraArgs: + - -v=4 + extraEnvs: {} + resources: {} + # If you do want to specify resources, uncomment the following lines, adjust them as necessary. + # and remove the curly braces after 'resources:'. +# limits: +# cpu: 1000m +# memory: 1000Mi +# requests: +# cpu: 100m +# memory: 100Mi + + deviceSplitCount: 10 + deviceMemoryScaling: 1 + deviceCoreScaling: 1 + # Node configuration for device plugin, Priority: externalConfigName > config > default config + nodeConfiguration: + # If you want to use a custom config.json, you can set the content here. + # If this is set, it will override the default config.json(An example is as follows). + config: | + { + "nodeconfig": [ + { + "name": "your-node-name", + "operatingmode": "hami-core", + "devicememoryscaling": 1, + "devicesplitcount": 10, + "migstrategy": "none", + "filterdevices": { + "uuid": [], + "index": [] + } + } + ] + } + # If you want to use an existing ConfigMap, you can set the name here. + # If this is set, the chart will not create the ConfigMap and will use the existing one. + externalConfigName: "" + # The runtime class name to be used by the device plugin, and added to the pod.spec.runtimeClassName of applications utilizing NVIDIA GPUs + runtimeClassName: "" + # Whether to create runtime class, name comes from runtimeClassName when it is set + createRuntimeClass: false + migStrategy: "none" + disablecorelimit: "false" + passDeviceSpecsEnabled: false + deviceListStrategy: "envvar" + nvidiaHookPath: null + nvidiaDriverRoot: null + gdrcopyEnabled: null + gdsEnabled: null + mofedEnabled: null + + extraArgs: + - -v=4 + extraEnvs: {} + service: + type: NodePort # Default type is NodePort, can be changed to ClusterIP + httpPort: 31992 + labels: {} + annotations: {} + + pluginPath: /var/lib/kubelet/device-plugins + libPath: /usr/local/vgpu + + podAnnotations: {} + nvidiaNodeSelector: + gpu: "on" + tolerations: [] + # The updateStrategy for DevicePlugin DaemonSet. + # If you want to update the DaemonSet by manual, set type as "OnDelete". + # We recommend use OnDelete update strategy because DevicePlugin pod restart will cause business pod restart, this behavior is destructive. + # Otherwise, you can use RollingUpdate update strategy to rolling update DevicePlugin pod. + updateStrategy: + type: RollingUpdate + rollingUpdate: + maxUnavailable: 1 + + resources: {} + # If you do want to specify resources, uncomment the following lines, adjust them as necessary. + # and remove the curly braces after 'resources:'. +# limits: +# cpu: 1000m +# memory: 1000Mi +# requests: +# cpu: 100m +# memory: 100Mi + +mockDevicePlugin: + enabled: false + image: + registry: "docker.io" + repository: "projecthami/mock-device-plugin" + tag: "1.0.1" + ## Specify a imagePullPolicy + ## Defaults to 'Always' if image tag is 'latest', else set to 'IfNotPresent' + ## ref: https://kubernetes.io/docs/user-guide/images/#pre-pulling-images + ## + pullPolicy: IfNotPresent + ## Optionally specify an array of imagePullSecrets. + ## Secrets must be manually created in the namespace. + ## Example: + ## pullSecrets: + ## - myRegistryKeySecretName + ## + pullSecrets: [] +devices: + amd: + customresources: + - amd.com/gpu + - amd.com/gpu-memory + awsneuron: + customresources: + - aws.amazon.com/neuron + - aws.amazon.com/neuroncore + kunlun: + enabled: true + customresources: + - kunlunxin.com/xpu + - kunlunxin.com/vxpu + - kunlunxin.com/vxpu-memory + enflame: + enabled: true + customresources: + - enflame.com/vgcu + - enflame.com/vgcu-percentage + - enflame.com/gcu + mthreads: + enabled: true + customresources: + - mthreads.com/vgpu + nvidia: + gpuCorePolicy: default + libCudaLogLevel: 1 + ascend: + enabled: false + image: "" + imagePullPolicy: IfNotPresent + extraArgs: [] + nodeSelector: + ascend: "on" + tolerations: [] + customresources: + - huawei.com/Ascend910A + - huawei.com/Ascend910A-memory + - huawei.com/Ascend910B2 + - huawei.com/Ascend910B2-memory + - huawei.com/Ascend910B3 + - huawei.com/Ascend910B3-memory + - huawei.com/Ascend910B4 + - huawei.com/Ascend910B4-memory + - huawei.com/Ascend910B4-1 + - huawei.com/Ascend910B4-1-memory + - huawei.com/Ascend310P + - huawei.com/Ascend310P-memory + iluvatar: + enabled: false + customresources: + - iluvatar.ai/BI-V100-vgpu + - iluvatar.ai/BI-V100.vCore + - iluvatar.ai/BI-V100.vMem + - iluvatar.ai/BI-V150-vgpu + - iluvatar.ai/BI-V150.vCore + - iluvatar.ai/BI-V150.vMem + - iluvatar.ai/MR-V100-vgpu + - iluvatar.ai/MR-V100.vCore + - iluvatar.ai/MR-V100.vMem + - iluvatar.ai/MR-V50-vgpu + - iluvatar.ai/MR-V50.vCore + - iluvatar.ai/MR-V50.vMem diff --git a/packages/system/hami/values.yaml b/packages/system/hami/values.yaml new file mode 100644 index 00000000..76e8d631 --- /dev/null +++ b/packages/system/hami/values.yaml @@ -0,0 +1,17 @@ +hami: + scheduler: + kubeScheduler: + image: + registry: registry.k8s.io + repository: kube-scheduler + devicePlugin: + runtimeClassName: nvidia + updateStrategy: + type: RollingUpdate + # See https://github.com/Project-HAMi/HAMi/blob/v2.8.1/deployments/charts/hami/values.yaml + # Each entry: {"name": "node-name", "operatingmode": "hami-core", "devicesplitcount": N, ...} + nodeConfiguration: + config: | + { + "nodeconfig": [] + } diff --git a/packages/system/kubernetes-rd/cozyrds/kubernetes.yaml b/packages/system/kubernetes-rd/cozyrds/kubernetes.yaml index 1631c86e..f81eaff0 100644 --- a/packages/system/kubernetes-rd/cozyrds/kubernetes.yaml +++ b/packages/system/kubernetes-rd/cozyrds/kubernetes.yaml @@ -22,7 +22,7 @@ spec: singular: kubernetes plural: kuberneteses openAPISchema: |- - {"title":"Chart Values","type":"object","properties":{"storageClass":{"description":"StorageClass used to store the data.","type":"string","default":"replicated"},"nodeGroups":{"description":"Worker nodes configuration map.","type":"object","default":{"md0":{"ephemeralStorage":"20Gi","gpus":[],"instanceType":"u1.medium","maxReplicas":10,"minReplicas":0,"resources":{},"roles":["ingress-nginx"]}},"additionalProperties":{"type":"object","required":["ephemeralStorage","instanceType","maxReplicas","minReplicas","resources"],"properties":{"ephemeralStorage":{"description":"Ephemeral storage size.","default":"20Gi","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"gpus":{"description":"List of GPUs to attach (NVIDIA driver requires at least 4 GiB RAM).","type":"array","items":{"type":"object","required":["name"],"properties":{"name":{"description":"Name of GPU, such as \"nvidia.com/AD102GL_L40S\".","type":"string"}}}},"instanceType":{"description":"Virtual machine instance type.","type":"string","default":"u1.medium"},"maxReplicas":{"description":"Maximum number of replicas.","type":"integer","default":10},"minReplicas":{"description":"Minimum number of replicas.","type":"integer","default":0},"resources":{"description":"CPU and memory resources for each worker node.","type":"object","properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"roles":{"description":"List of node roles.","type":"array","items":{"type":"string"}}}}},"version":{"description":"Kubernetes major.minor version to deploy","type":"string","default":"v1.35","enum":["v1.35","v1.34","v1.33","v1.32","v1.31","v1.30"]},"host":{"description":"External hostname for Kubernetes cluster. Defaults to `.` if empty.","type":"string","default":""},"addons":{"description":"Cluster addons configuration.","type":"object","default":{},"required":["certManager","cilium","coredns","fluxcd","gatewayAPI","gpuOperator","ingressNginx","monitoringAgents","velero","verticalPodAutoscaler"],"properties":{"certManager":{"description":"Cert-manager addon.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable cert-manager.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"cilium":{"description":"Cilium CNI plugin.","type":"object","default":{},"required":["valuesOverride"],"properties":{"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"coredns":{"description":"CoreDNS addon.","type":"object","default":{},"required":["valuesOverride"],"properties":{"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"fluxcd":{"description":"FluxCD GitOps operator.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable FluxCD.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"gatewayAPI":{"description":"Gateway API addon.","type":"object","default":{},"required":["enabled"],"properties":{"enabled":{"description":"Enable Gateway API.","type":"boolean","default":false}}},"gpuOperator":{"description":"NVIDIA GPU Operator.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable GPU Operator.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"ingressNginx":{"description":"Ingress-NGINX controller.","type":"object","default":{},"required":["enabled","exposeMethod","valuesOverride"],"properties":{"enabled":{"description":"Enable the controller (requires nodes labeled `ingress-nginx`).","type":"boolean","default":false},"exposeMethod":{"description":"Method to expose the controller. Allowed values: `Proxied`, `LoadBalancer`.","type":"string","default":"Proxied","enum":["Proxied","LoadBalancer"]},"hosts":{"description":"Domains routed to this tenant cluster when `exposeMethod` is `Proxied`.","type":"array","default":[],"items":{"type":"string"}},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"monitoringAgents":{"description":"Monitoring agents.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable monitoring agents.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"velero":{"description":"Velero backup/restore addon.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable Velero.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"verticalPodAutoscaler":{"description":"Vertical Pod Autoscaler.","type":"object","default":{},"required":["valuesOverride"],"properties":{"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}}}},"controlPlane":{"description":"Kubernetes control-plane configuration.","type":"object","default":{},"required":["apiServer","controllerManager","konnectivity","replicas","scheduler"],"properties":{"apiServer":{"description":"API Server configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for API Server.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"large","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}},"controllerManager":{"description":"Controller Manager configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for Controller Manager.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"micro","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}},"konnectivity":{"description":"Konnectivity configuration.","type":"object","default":{},"required":["server"],"properties":{"server":{"description":"Konnectivity Server configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for Konnectivity.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"micro","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}}}},"replicas":{"description":"Number of control-plane replicas.","type":"integer","default":2},"scheduler":{"description":"Scheduler configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for Scheduler.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"micro","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}}}},"images":{"description":"Optional image overrides for air-gapped or rate-limited registries.","type":"object","default":{},"properties":{"waitForKubeconfig":{"description":"Image used by the wait-for-kubeconfig init container. Empty falls back to images/busybox.tag.","type":"string","default":""}}}}} + {"title":"Chart Values","type":"object","properties":{"storageClass":{"description":"StorageClass used to store the data.","type":"string","default":"replicated"},"nodeGroups":{"description":"Worker nodes configuration map.","type":"object","default":{"md0":{"ephemeralStorage":"20Gi","gpus":[],"instanceType":"u1.medium","maxReplicas":10,"minReplicas":0,"resources":{},"roles":["ingress-nginx"]}},"additionalProperties":{"type":"object","required":["ephemeralStorage","instanceType","maxReplicas","minReplicas","resources"],"properties":{"ephemeralStorage":{"description":"Ephemeral storage size.","default":"20Gi","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"gpus":{"description":"List of GPUs to attach (NVIDIA driver requires at least 4 GiB RAM).","type":"array","items":{"type":"object","required":["name"],"properties":{"name":{"description":"Name of GPU, such as \"nvidia.com/AD102GL_L40S\".","type":"string"}}}},"instanceType":{"description":"Virtual machine instance type.","type":"string","default":"u1.medium"},"maxReplicas":{"description":"Maximum number of replicas.","type":"integer","default":10},"minReplicas":{"description":"Minimum number of replicas.","type":"integer","default":0},"resources":{"description":"CPU and memory resources for each worker node.","type":"object","properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"roles":{"description":"List of node roles.","type":"array","items":{"type":"string"}}}}},"version":{"description":"Kubernetes major.minor version to deploy","type":"string","default":"v1.35","enum":["v1.35","v1.34","v1.33","v1.32","v1.31","v1.30"]},"host":{"description":"External hostname for Kubernetes cluster. Defaults to `.` if empty.","type":"string","default":""},"addons":{"description":"Cluster addons configuration.","type":"object","default":{},"required":["certManager","cilium","coredns","fluxcd","gatewayAPI","gpuOperator","hami","ingressNginx","monitoringAgents","velero","verticalPodAutoscaler"],"properties":{"certManager":{"description":"Cert-manager addon.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable cert-manager.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"cilium":{"description":"Cilium CNI plugin.","type":"object","default":{},"required":["valuesOverride"],"properties":{"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"coredns":{"description":"CoreDNS addon.","type":"object","default":{},"required":["valuesOverride"],"properties":{"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"fluxcd":{"description":"FluxCD GitOps operator.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable FluxCD.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"gatewayAPI":{"description":"Gateway API addon.","type":"object","default":{},"required":["enabled"],"properties":{"enabled":{"description":"Enable Gateway API.","type":"boolean","default":false}}},"gpuOperator":{"description":"NVIDIA GPU Operator.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable GPU Operator.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"hami":{"description":"HAMi GPU virtualization middleware.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable HAMi (requires GPU Operator).","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"ingressNginx":{"description":"Ingress-NGINX controller.","type":"object","default":{},"required":["enabled","exposeMethod","valuesOverride"],"properties":{"enabled":{"description":"Enable the controller (requires nodes labeled `ingress-nginx`).","type":"boolean","default":false},"exposeMethod":{"description":"Method to expose the controller. Allowed values: `Proxied`, `LoadBalancer`.","type":"string","default":"Proxied","enum":["Proxied","LoadBalancer"]},"hosts":{"description":"Domains routed to this tenant cluster when `exposeMethod` is `Proxied`.","type":"array","default":[],"items":{"type":"string"}},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"monitoringAgents":{"description":"Monitoring agents.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable monitoring agents.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"velero":{"description":"Velero backup/restore addon.","type":"object","default":{},"required":["enabled","valuesOverride"],"properties":{"enabled":{"description":"Enable Velero.","type":"boolean","default":false},"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}},"verticalPodAutoscaler":{"description":"Vertical Pod Autoscaler.","type":"object","default":{},"required":["valuesOverride"],"properties":{"valuesOverride":{"description":"Custom Helm values overrides.","type":"object","default":{},"x-kubernetes-preserve-unknown-fields":true}}}}},"controlPlane":{"description":"Kubernetes control-plane configuration.","type":"object","default":{},"required":["apiServer","controllerManager","konnectivity","replicas","scheduler"],"properties":{"apiServer":{"description":"API Server configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for API Server.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"large","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}},"controllerManager":{"description":"Controller Manager configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for Controller Manager.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"micro","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}},"konnectivity":{"description":"Konnectivity configuration.","type":"object","default":{},"required":["server"],"properties":{"server":{"description":"Konnectivity Server configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for Konnectivity.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"micro","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}}}},"replicas":{"description":"Number of control-plane replicas.","type":"integer","default":2},"scheduler":{"description":"Scheduler configuration.","type":"object","default":{},"required":["resources","resourcesPreset"],"properties":{"resources":{"description":"CPU and memory resources for Scheduler.","type":"object","default":{},"properties":{"cpu":{"description":"CPU available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true},"memory":{"description":"Memory (RAM) available.","pattern":"^(\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\\+|-)?(([0-9]+(\\.[0-9]*)?)|(\\.[0-9]+))))?$","anyOf":[{"type":"integer"},{"type":"string"}],"x-kubernetes-int-or-string":true}}},"resourcesPreset":{"description":"Preset if `resources` omitted.","type":"string","default":"micro","enum":["nano","micro","small","medium","large","xlarge","2xlarge"]}}}}},"images":{"description":"Optional image overrides for air-gapped or rate-limited registries.","type":"object","default":{},"properties":{"waitForKubeconfig":{"description":"Image used by the wait-for-kubeconfig init container. Empty falls back to images/busybox.tag.","type":"string","default":""}}}}} release: prefix: kubernetes- labels: @@ -40,7 +40,7 @@ spec: tags: - compute icon: <svg width="144" height="144" viewBox="0 0 144 144" fill="none" xmlns="http://www.w3.org/2000/svg">
<rect width="144" height="144" rx="24" fill="url(#paint0_linear_681_2845)"/>
<path d="M71.9968 19C70.3039 19.0002 68.9312 20.5322 68.9314 22.4221C68.9314 22.4511 68.9373 22.4788 68.9379 22.5076C68.9354 22.7644 68.9231 23.0737 68.9314 23.2973C68.9717 24.3873 69.2082 25.2216 69.3506 26.2258C69.6084 28.3752 69.8245 30.1569 69.6912 31.813C69.5615 32.4375 69.1037 33.0086 68.6956 33.4056L68.6235 34.7086C66.7839 34.8617 64.9319 35.1421 63.0821 35.5641C55.1226 37.3798 48.2695 41.4991 43.052 47.0609C42.7134 46.8288 42.1211 46.4019 41.945 46.2712C41.3977 46.3454 40.8445 46.5151 40.1241 46.0935C38.7522 45.1657 37.5028 43.8851 35.991 42.3424C35.2982 41.6044 34.7966 40.9018 33.9735 40.1904C33.7866 40.0289 33.5014 39.8104 33.2923 39.6442C32.6489 39.1288 31.89 38.86 31.157 38.8348C30.2147 38.8024 29.3075 39.1725 28.7138 39.9206C27.6584 41.2506 27.9963 43.2833 29.4671 44.4614C29.482 44.4734 29.4979 44.4827 29.5129 44.4943C29.715 44.6589 29.9625 44.8698 30.1483 45.0076C31.0217 45.6555 31.8195 45.9872 32.6897 46.5015C34.5231 47.6391 36.043 48.5823 37.2486 49.7196C37.7194 50.2237 37.8016 51.1122 37.8643 51.4964L38.8468 52.3782C33.5872 60.3308 31.153 70.1539 32.5915 80.1627L31.3077 80.5378C30.9693 80.9768 30.4912 81.6676 29.9911 81.8738C28.4138 82.3729 26.6387 82.5562 24.4956 82.7819C23.4894 82.866 22.6213 82.8158 21.5546 83.0188C21.3198 83.0635 20.9927 83.1491 20.7358 83.2097C20.7269 83.2116 20.7186 83.2142 20.7096 83.2162C20.6956 83.2195 20.6772 83.2263 20.6638 83.2294C18.857 83.668 17.6963 85.3365 18.0699 86.9805C18.4437 88.6248 20.2086 89.6248 22.0262 89.2312C22.0393 89.2282 22.0584 89.2277 22.072 89.2246C22.0926 89.2199 22.1106 89.2099 22.131 89.2049C22.3844 89.149 22.7019 89.0868 22.9236 89.0272C23.9723 88.7451 24.7318 88.3306 25.6746 87.9677C27.7029 87.2368 29.3828 86.6262 31.0195 86.3883C31.703 86.3345 32.4232 86.812 32.7814 87.0134L34.1177 86.7831C37.1926 96.3613 43.6366 104.103 51.7963 108.961L51.2396 110.303C51.4403 110.824 51.6616 111.53 51.5121 112.045C50.9171 113.595 49.898 115.231 48.7374 117.055C48.1755 117.898 47.6004 118.552 47.0934 119.516C46.972 119.747 46.8175 120.102 46.7004 120.346C45.9125 122.039 46.4904 123.99 48.0038 124.722C49.5268 125.459 51.4171 124.682 52.2352 122.985C52.2364 122.982 52.2406 122.98 52.2417 122.978C52.2426 122.976 52.2409 122.973 52.2417 122.971C52.3582 122.731 52.5233 122.415 52.6216 122.188C53.056 121.189 53.2005 120.332 53.5059 119.365C54.317 117.318 54.7626 115.17 55.8791 113.832C56.1849 113.466 56.6833 113.325 57.2001 113.186L57.8944 111.922C65.008 114.665 72.9705 115.402 80.9245 113.587C82.7391 113.173 84.4908 112.637 86.1843 111.994C86.3794 112.342 86.742 113.011 86.8393 113.179C87.3644 113.351 87.9377 113.439 88.4047 114.133C89.2401 115.567 89.8114 117.263 90.5073 119.312C90.8128 120.279 90.9638 121.136 91.3981 122.136C91.4971 122.363 91.6614 122.684 91.778 122.925C92.5944 124.628 94.4907 125.407 96.0159 124.669C97.5292 123.937 98.1077 121.986 97.3194 120.293C97.2023 120.049 97.0412 119.695 96.9198 119.464C96.4127 118.499 95.8377 117.852 95.2758 117.009C94.1152 115.185 93.1526 113.67 92.5575 112.12C92.3087 111.32 92.5995 110.823 92.7933 110.303C92.6772 110.17 92.4288 109.414 92.2824 109.059C100.762 104.029 107.017 95.9985 109.955 86.7239C110.351 86.7865 111.041 86.9091 111.265 86.9542C111.726 86.6487 112.149 86.2501 112.981 86.3159C114.617 86.5537 116.297 87.1645 118.326 87.8953C119.268 88.2581 120.028 88.6793 121.077 88.9614C121.298 89.021 121.616 89.0766 121.869 89.1325C121.889 89.1375 121.908 89.1475 121.928 89.1522C121.942 89.1553 121.961 89.1558 121.974 89.1588C123.792 89.552 125.557 88.5526 125.93 86.9081C126.303 85.2641 125.143 83.5952 123.336 83.157C123.074 83.097 122.701 82.995 122.446 82.9465C121.379 82.7435 120.511 82.7935 119.505 82.7095C117.361 82.4839 115.586 82.3004 114.009 81.8014C113.366 81.5507 112.908 80.7819 112.686 80.4655L111.448 80.1035C112.09 75.438 111.917 70.5825 110.806 65.7243C109.685 60.8208 107.704 56.3361 105.062 52.3848C105.379 52.0948 105.979 51.5612 106.149 51.4043C106.199 50.8517 106.156 50.2722 106.725 49.6603C107.931 48.523 109.451 47.5799 111.284 46.4423C112.154 45.9279 112.959 45.5964 113.832 44.9484C114.03 44.8019 114.299 44.5699 114.507 44.4022C115.977 43.2237 116.316 41.1911 115.26 39.8614C114.204 38.5317 112.159 38.4065 110.688 39.585C110.479 39.7516 110.195 39.9688 110.007 40.1312C109.184 40.8426 108.676 41.5452 107.983 42.2832C106.471 43.8259 105.222 45.113 103.85 46.0409C103.255 46.3885 102.385 46.2682 101.99 46.2449L100.824 47.0806C94.1753 40.0763 85.1235 35.5982 75.3766 34.7283C75.3494 34.3179 75.3137 33.5761 75.3046 33.3529C74.9056 32.9693 74.4235 32.6418 74.3024 31.813C74.1691 30.1569 74.3917 28.3752 74.6496 26.2258C74.7919 25.2216 75.0284 24.3873 75.0688 23.2973C75.078 23.0495 75.0632 22.69 75.0622 22.4221C75.062 20.5322 73.6898 18.9998 71.9968 19ZM68.1585 42.8886L67.248 59.0447L67.1825 59.0776C67.1214 60.5229 65.9375 61.677 64.4839 61.677C63.8884 61.677 63.3388 61.4849 62.8922 61.1571L62.866 61.1703L49.6807 51.7794C53.7331 47.7759 58.9164 44.8172 64.89 43.4546C65.9812 43.2056 67.0719 43.0209 68.1585 42.8886ZM75.8417 42.8886C82.8159 43.7504 89.2657 46.9232 94.2081 51.786L81.108 61.1176L81.0621 61.0979C79.8994 61.9512 78.2611 61.7394 77.3548 60.5978C76.9835 60.1301 76.7887 59.5801 76.7653 59.0249L76.7522 59.0184L75.8417 42.8886ZM44.8991 57.814L56.9382 68.633L56.9251 68.6988C58.0117 69.6479 58.172 71.2949 57.2657 72.4368C56.8944 72.9045 56.3975 73.2182 55.8639 73.3647L55.8508 73.4173L40.4188 77.8923C39.6334 70.6765 41.3261 63.6621 44.8991 57.814ZM99.0094 57.8206C100.798 60.7336 102.153 63.9871 102.959 67.5143C103.756 70.9991 103.956 74.4778 103.627 77.8397L88.1166 73.3515L88.1035 73.2857C86.7145 72.9043 85.8609 71.4848 86.1843 70.0611C86.3168 69.4778 86.6249 68.9844 87.0423 68.6198L87.0358 68.5869L99.0094 57.8206ZM69.5274 69.4688H74.4596L77.5251 73.3186L76.4247 78.1226L71.9968 80.2614L67.5558 78.1161L66.4554 73.312L69.5274 69.4688ZM85.3393 82.6437C85.5489 82.6331 85.7576 82.652 85.9616 82.6898L85.9878 82.6569L101.95 85.3682C99.6142 91.9624 95.144 97.675 89.1711 101.498L82.9747 86.4606L82.9944 86.4343C82.4252 85.1055 82.9948 83.5472 84.3044 82.9135C84.6397 82.7513 84.99 82.6614 85.3393 82.6437ZM58.5298 82.7095C59.748 82.7267 60.8406 83.5761 61.1237 84.822C61.2562 85.4052 61.1917 85.9831 60.973 86.4935L61.0189 86.5527L54.888 101.439C49.1559 97.7432 44.5904 92.2099 42.1481 85.4208L57.9731 82.7227L57.9993 82.7556C58.1763 82.7229 58.3558 82.7071 58.5298 82.7095ZM71.8986 89.2312C72.3229 89.2155 72.7534 89.303 73.1627 89.501C73.6992 89.7606 74.1136 90.1692 74.3745 90.6592H74.4334L82.2346 104.821C81.2221 105.162 80.1813 105.454 79.1167 105.697C73.1505 107.058 67.2032 106.645 61.818 104.802L69.5995 90.6658H69.6126C70.0795 89.7888 70.965 89.2656 71.8986 89.2312Z" fill="white" stroke="white" stroke-width="0.25"/>
<defs>
<linearGradient id="paint0_linear_681_2845" x1="10" y1="15.5" x2="144" y2="131.5" gradientUnits="userSpaceOnUse">
<stop stop-color="#4D87FF"/>
<stop offset="1" stop-color="#0547D0"/>
</linearGradient>
</defs>
</svg>
 - keysOrder: [["apiVersion"], ["appVersion"], ["kind"], ["metadata"], ["metadata", "name"], ["spec", "storageClass"], ["spec", "nodeGroups"], ["spec", "nodeGroups", "md0"], ["spec", "nodeGroups", "md0", "minReplicas"], ["spec", "nodeGroups", "md0", "maxReplicas"], ["spec", "nodeGroups", "md0", "instanceType"], ["spec", "nodeGroups", "md0", "ephemeralStorage"], ["spec", "nodeGroups", "md0", "roles"], ["spec", "nodeGroups", "md0", "resources"], ["spec", "nodeGroups", "md0", "gpus"], ["spec", "version"], ["spec", "host"], ["spec", "addons"], ["spec", "addons", "certManager"], ["spec", "addons", "certManager", "enabled"], ["spec", "addons", "certManager", "valuesOverride"], ["spec", "addons", "cilium"], ["spec", "addons", "cilium", "valuesOverride"], ["spec", "addons", "gatewayAPI"], ["spec", "addons", "gatewayAPI", "enabled"], ["spec", "addons", "ingressNginx"], ["spec", "addons", "ingressNginx", "enabled"], ["spec", "addons", "ingressNginx", "exposeMethod"], ["spec", "addons", "ingressNginx", "hosts"], ["spec", "addons", "ingressNginx", "valuesOverride"], ["spec", "addons", "gpuOperator"], ["spec", "addons", "gpuOperator", "enabled"], ["spec", "addons", "gpuOperator", "valuesOverride"], ["spec", "addons", "fluxcd"], ["spec", "addons", "fluxcd", "enabled"], ["spec", "addons", "fluxcd", "valuesOverride"], ["spec", "addons", "monitoringAgents"], ["spec", "addons", "monitoringAgents", "enabled"], ["spec", "addons", "monitoringAgents", "valuesOverride"], ["spec", "addons", "verticalPodAutoscaler"], ["spec", "addons", "verticalPodAutoscaler", "valuesOverride"], ["spec", "addons", "velero"], ["spec", "addons", "velero", "enabled"], ["spec", "addons", "velero", "valuesOverride"], ["spec", "addons", "coredns"], ["spec", "addons", "coredns", "valuesOverride"], ["spec", "controlPlane"], ["spec", "controlPlane", "replicas"], ["spec", "controlPlane", "apiServer"], ["spec", "controlPlane", "apiServer", "resources"], ["spec", "controlPlane", "apiServer", "resourcesPreset"], ["spec", "controlPlane", "controllerManager"], ["spec", "controlPlane", "controllerManager", "resources"], ["spec", "controlPlane", "controllerManager", "resourcesPreset"], ["spec", "controlPlane", "scheduler"], ["spec", "controlPlane", "scheduler", "resources"], ["spec", "controlPlane", "scheduler", "resourcesPreset"], ["spec", "controlPlane", "konnectivity"], ["spec", "controlPlane", "konnectivity", "server"], ["spec", "controlPlane", "konnectivity", "server", "resources"], ["spec", "controlPlane", "konnectivity", "server", "resourcesPreset"], ["spec", "images"], ["spec", "images", "waitForKubeconfig"]] + keysOrder: [["apiVersion"], ["appVersion"], ["kind"], ["metadata"], ["metadata", "name"], ["spec", "storageClass"], ["spec", "nodeGroups"], ["spec", "nodeGroups", "md0"], ["spec", "nodeGroups", "md0", "minReplicas"], ["spec", "nodeGroups", "md0", "maxReplicas"], ["spec", "nodeGroups", "md0", "instanceType"], ["spec", "nodeGroups", "md0", "ephemeralStorage"], ["spec", "nodeGroups", "md0", "roles"], ["spec", "nodeGroups", "md0", "resources"], ["spec", "nodeGroups", "md0", "gpus"], ["spec", "version"], ["spec", "host"], ["spec", "addons"], ["spec", "addons", "certManager"], ["spec", "addons", "certManager", "enabled"], ["spec", "addons", "certManager", "valuesOverride"], ["spec", "addons", "cilium"], ["spec", "addons", "cilium", "valuesOverride"], ["spec", "addons", "gatewayAPI"], ["spec", "addons", "gatewayAPI", "enabled"], ["spec", "addons", "ingressNginx"], ["spec", "addons", "ingressNginx", "enabled"], ["spec", "addons", "ingressNginx", "exposeMethod"], ["spec", "addons", "ingressNginx", "hosts"], ["spec", "addons", "ingressNginx", "valuesOverride"], ["spec", "addons", "gpuOperator"], ["spec", "addons", "gpuOperator", "enabled"], ["spec", "addons", "gpuOperator", "valuesOverride"], ["spec", "addons", "hami"], ["spec", "addons", "hami", "enabled"], ["spec", "addons", "hami", "valuesOverride"], ["spec", "addons", "fluxcd"], ["spec", "addons", "fluxcd", "enabled"], ["spec", "addons", "fluxcd", "valuesOverride"], ["spec", "addons", "monitoringAgents"], ["spec", "addons", "monitoringAgents", "enabled"], ["spec", "addons", "monitoringAgents", "valuesOverride"], ["spec", "addons", "verticalPodAutoscaler"], ["spec", "addons", "verticalPodAutoscaler", "valuesOverride"], ["spec", "addons", "velero"], ["spec", "addons", "velero", "enabled"], ["spec", "addons", "velero", "valuesOverride"], ["spec", "addons", "coredns"], ["spec", "addons", "coredns", "valuesOverride"], ["spec", "controlPlane"], ["spec", "controlPlane", "replicas"], ["spec", "controlPlane", "apiServer"], ["spec", "controlPlane", "apiServer", "resources"], ["spec", "controlPlane", "apiServer", "resourcesPreset"], ["spec", "controlPlane", "controllerManager"], ["spec", "controlPlane", "controllerManager", "resources"], ["spec", "controlPlane", "controllerManager", "resourcesPreset"], ["spec", "controlPlane", "scheduler"], ["spec", "controlPlane", "scheduler", "resources"], ["spec", "controlPlane", "scheduler", "resourcesPreset"], ["spec", "controlPlane", "konnectivity"], ["spec", "controlPlane", "konnectivity", "server"], ["spec", "controlPlane", "konnectivity", "server", "resources"], ["spec", "controlPlane", "konnectivity", "server", "resourcesPreset"], ["spec", "images"], ["spec", "images", "waitForKubeconfig"]] secrets: exclude: [] include: