Merge pull request #168 from nebius/dev

Release 1.15.1

.github/workflows/github_release.yml

@@ -63,7 +63,7 @@ jobs:
 
       - name: Generate changelog
         id: changelog
-        uses: mikepenz/release-changelog-builder-action@f3fc77b47b74e78971fffecb2102ae6eac9a44d6 # v5
+        uses: mikepenz/release-changelog-builder-action@a57c1b7c90e56d9c8b26a6ed5d1eed159369e117 # v5
         with:
           mode: "PR"
           fromTag: ${{ needs.tag.outputs.previous-tag }}
@@ -117,7 +117,7 @@ jobs:
           token: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Create GitHub Release with changelog
-        uses: softprops/action-gh-release@c062e08bd532815e2082a85e87e3ef29c3e6d191 # v2.0.8
+        uses: softprops/action-gh-release@e7a8f85e1c67a31e6ed99a94b41bd0b71bbee6b8 # v2.0.9
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         with:

.github/workflows/one_job.yml

@@ -52,9 +52,6 @@ jobs:
         with:
           egress-policy: audit
 
-      - name: Debug vars
-        run: echo "UNSTABLE - is ${{ needs.pre-build.outputs.unstable }}"
-
       - name: Checkout repository
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
 
@@ -63,14 +60,19 @@ jobs:
         with:
             go-version-file: 'go.mod'
 
+      - name: Debug vars
+        run: |
+          echo "UNSTABLE - is ${{ needs.pre-build.outputs.unstable }}"
+          make get-version UNSTABLE=${{ needs.pre-build.outputs.unstable }}
+
       - name: Check if version synced
         run: make test-version-sync
 
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@c47758b77c9736f4b2ef4073d4d51994fabfe349 # v3.7.1
 
       - name: Log in to the Github Container registry
-        uses: docker/login-action@06895751d15a223ec091bea144ad5c7f50d228d0
+        uses: docker/login-action@7ca345011ac4304463197fac0e56eab1bc7e6af0
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
@@ -108,6 +110,9 @@ jobs:
 
           make docker-build UNSTABLE="${UNSTABLE}" IMAGE_NAME=exporter DOCKERFILE=exporter/exporter.dockerfile
           make docker-push  UNSTABLE="${UNSTABLE}" IMAGE_NAME=exporter
+          
+          make docker-build UNSTABLE="${UNSTABLE}" IMAGE_NAME=slurmrestd DOCKERFILE=restd/slurmrestd.dockerfile
+          make docker-push  UNSTABLE="${UNSTABLE}" IMAGE_NAME=slurmrestd
 
           echo "Common images were built"
 

Makefile

@@ -43,17 +43,12 @@ VALUES_VERSION 		  = $(shell $(YQ) '.images.slurmctld' helm/slurm-cluster/values
 OPERATOR_IMAGE_TAG  = $(VERSION)
 
 ifeq ($(shell uname), Darwin)
-    SHA_CMD = shasum -a 256
     SED_COMMAND = sed -i '' -e
-    USER_MAIL					= $(shell git config user.email)
 else
-    SHA_CMD = sha256sum
     SED_COMMAND = sed -i -e
-    USER_MAIL					= $(shell git log -1 --pretty=format:'%ae')
 endif
 ifeq ($(UNSTABLE), true)
-	USER_MAIL					= $(shell git log -1 --pretty=format:'%ae')
-    SHORT_SHA 					= $(shell echo -n "$(USER_MAIL)-$(VERSION)" | $(SHA_CMD) | cut -c1-8)
+    SHORT_SHA 					= $(shell git rev-parse --short=8 HEAD)
     OPERATOR_IMAGE_TAG  		= $(VERSION)-$(SHORT_SHA)
     IMAGE_VERSION		  		= $(VERSION)-$(UBUNTU_VERSION)-slurm$(SLURM_VERSION)-$(SHORT_SHA)
     IMAGE_REPO			  		= $(NEBIUS_REPO)-unstable
@@ -177,6 +172,7 @@ sync-version: yq ## Sync versions from file
 	@echo 'Syncing helm/slurm-cluster/values.yaml'
 	@$(YQ) -i ".images.ncclBenchmark = \"$(IMAGE_REPO)/nccl_benchmark:$(IMAGE_VERSION)\"" "helm/slurm-cluster/values.yaml"
 	@$(YQ) -i ".images.slurmctld = \"$(IMAGE_REPO)/controller_slurmctld:$(IMAGE_VERSION)\"" "helm/slurm-cluster/values.yaml"
+	@$(YQ) -i ".images.slurmrestd = \"$(IMAGE_REPO)/slurmrestd:$(IMAGE_VERSION)\"" "helm/slurm-cluster/values.yaml"
 	@$(YQ) -i ".images.slurmdbd = \"$(IMAGE_REPO)/controller_slurmdbd:$(IMAGE_VERSION)\"" "helm/slurm-cluster/values.yaml"
 	@$(YQ) -i ".images.slurmd = \"$(IMAGE_REPO)/worker_slurmd:$(IMAGE_VERSION)\"" "helm/slurm-cluster/values.yaml"
 	@$(YQ) -i ".images.sshd = \"$(IMAGE_REPO)/login_sshd:$(IMAGE_VERSION)\"" "helm/slurm-cluster/values.yaml"

README.md

@@ -140,10 +140,10 @@ Everything specific to Nebius AI is contained in a separate repository:
 ### Other clouds and on-premises
 > [!IMPORTANT]
 > When using the soperator, it is important that the CNI supports preserving the client source IP.
-> Therefore, if kube-proxy is configured in IPVS mode, or if you're using CNIplugins like kube-router or Antrea Proxy,
+> Therefore, if kube-proxy is configured in IPVS mode, or if you're using CNI plugins like kube-router or Antrea Proxy,
 the operator will not work.
 > This operator has been tested with
-the[Cilium network plugin](https://kubernetes.io/docs/concepts/extend-kubernetes/compute-storage-net/network-plugins/)
+the [Cilium network plugin](https://kubernetes.io/docs/concepts/extend-kubernetes/compute-storage-net/network-plugins/)
 > running in
 [kube-proxy replacement mode](https://docs.cilium.io/en/stable/network/kubernetes/kubeproxy-free/#kubernetes-without-kube-proxy).
 

VERSION

@@ -1 +1 @@
-1.14.14
+1.15.1

api/v1/slurmcluster_types.go

@@ -110,6 +110,13 @@ type PopulateJail struct {
 	// +kubebuilder:validation:Required
 	Image string `json:"image"`
 
+	// ImagePullPolicy defines the image pull policy
+	//
+	// +kubebuilder:validation:Enum=Always;Never;IfNotPresent
+	// +kubebuilder:validation:Optional
+	// +kubebuilder:default="IfNotPresent"
+	ImagePullPolicy corev1.PullPolicy `json:"imagePullPolicy,omitempty"`
+
 	// K8sNodeFilterName defines the Kubernetes node filter name associated with the Slurm node.
 	// Must correspond to the name of one of [K8sNodeFilter]
 	//
@@ -175,6 +182,13 @@ type NCCLBenchmark struct {
 	// +kubebuilder:validation:Required
 	Image string `json:"image"`
 
+	// ImagePullPolicy defines the image pull policy
+	//
+	// +kubebuilder:validation:Enum=Always;Never;IfNotPresent
+	// +kubebuilder:validation:Optional
+	// +kubebuilder:default="IfNotPresent"
+	ImagePullPolicy corev1.PullPolicy `json:"imagePullPolicy,omitempty"`
+
 	// NCCLArguments define nccl settings
 	//
 	// +kubebuilder:validation:Optional
@@ -316,6 +330,24 @@ type SlurmNodes struct {
 	// TODO: Making exporter optional requires SlurmNode.K8sNodeFilterName to be optional.
 	// +kubebuilder:validation:Required
 	Exporter SlurmExporter `json:"exporter"`
+
+	Rest SlurmRest `json:"rest"`
+}
+
+// SlurmRest represents the Slur REST API configuration
+type SlurmRest struct {
+	SlurmNode `json:",inline"`
+
+	// Enabled defines whether the SlurmRest is enabled
+	//
+	// +kubebuilder:validation:Optional
+	// +kubebuilder:default=false
+	Enabled bool `json:"enabled,omitempty"`
+
+	// SlurmRestNode represents the Slurm REST API daemon configuration
+	//
+	// +kubebuilder:validation:Optional
+	SlurmRestNode NodeContainer `json:"rest,omitempty"`
 }
 
 // SlurmNodeAccounting represents the Slurm accounting configuration
@@ -666,7 +698,7 @@ type ExporterContainer struct {
 	PodTemplateNameRef *string `json:"podTemplateNameRef,omitempty"`
 }
 
-// SlurmExporterVolumes define the volumes for the Slurm controller node
+// SlurmExporterVolumes define the volumes for the Slurm exporter node
 type SlurmExporterVolumes struct {
 	// Jail represents the jail data volume configuration
 	//
@@ -693,6 +725,13 @@ type NodeContainer struct {
 	// +kubebuilder:validation:Required
 	Image string `json:"image"`
 
+	// ImagePullPolicy defines the image pull policy
+	//
+	// +kubebuilder:validation:Enum=Always;Never;IfNotPresent
+	// +kubebuilder:validation:Optional
+	// +kubebuilder:default="IfNotPresent"
+	ImagePullPolicy corev1.PullPolicy `json:"imagePullPolicy,omitempty"`
+
 	// Port defines the port the container exposes
 	//
 	// +kubebuilder:validation:Optional
@@ -709,6 +748,12 @@ type NodeContainer struct {
 	//
 	// +kubebuilder:validation:Optional
 	SecurityLimitsConfig string `json:"securityLimitsConfig,omitempty"`
+
+	// AppArmorProfile defines the AppArmor profile for the Slurm worker node
+	//
+	// +kubebuilder:validation:Optional
+	// +kubebuilder:default="unconfined"
+	AppArmorProfile string `json:"appArmorProfile,omitempty"`
 }
 
 // NodeVolume defines the configuration for a node volume.