Merge release v0.1.12

Release v0.1.12

.github/workflows/main.yml

@@ -43,8 +43,9 @@ jobs:
     - name: "Verify auto-generated files"
       run: |
           make manifests generate
+          git status
           if [[ $(git status -s | wc -l) -gt 0 ]]; then \
-              git status; exit 1; \
+              exit 1; \
           fi
 
     - name: "Docker metadata"

.vscode/launch.json

@@ -27,6 +27,18 @@
                 "--socket=/tmp/nnf.sock"
             ]
         },
+        {
+            "name": "Debug Copy-Offload server",
+            "type": "go",
+            "request": "launch",
+            "mode": "debug",
+            "program": "${workspaceFolder}/daemons/copy-offload/cmd",
+            "env": {"NNF_NODE_NAME": "rabbit-node-1"},
+            "args" :[
+                "--port=8080",
+                "--mock",
+            ]
+        },
         {
             "name": "Launch file",
             "type": "go",

Dockerfile

@@ -18,13 +18,10 @@
 # These ARGs must be before the first FROM. This allows them to be valid for
 # use in FROM instructions.
 ARG NNFMFU_TAG_BASE=ghcr.io/nearnodeflash/nnf-mfu
-ARG NNFMFU_VERSION=0.1.3
+ARG NNFMFU_VERSION=0.1.4
 
 # Build the manager binary
-FROM golang:1.21-alpine AS builder
-
-ARG TARGETARCH
-ARG TARGETOS
+FROM golang:1.21-alpine AS builder_setup
 
 WORKDIR /workspace
 # Copy the Go Modules manifests
@@ -40,16 +37,39 @@ COPY cmd/ cmd/
 COPY api/ api/
 COPY internal/ internal/
 
+###############################################################################
+FROM builder_setup AS builder
+
+ARG TARGETARCH
+ARG TARGETOS
+
 # Build
 # the GOARCH has a default value to allow the binary be built according to the host where the command
 # was called. For example, if we call make docker-build in a local env which has the Apple Silicon M1 SO
 # the docker BUILDPLATFORM arg will be linux/arm64 when for Apple x86 it will be linux/amd64. Therefore,
 # by leaving it empty we can ensure that the container and binary shipped on it will have the same platform.
 RUN CGO_ENABLED=0 GOOS=${TARGETOS:-linux} GOARCH=${TARGETARCH} go build -a -o manager cmd/main.go
 
+###############################################################################
+FROM builder_setup AS copy_offload_builder
+
+ARG TARGETARCH
+ARG TARGETOS
+
+COPY daemons/copy-offload/ daemons/copy-offload/
+
+# Build
+# the GOARCH has a default value to allow the binary be built according to the host where the command
+# was called. For example, if we call make docker-build in a local env which has the Apple Silicon M1 SO
+# the docker BUILDPLATFORM arg will be linux/arm64 when for Apple x86 it will be linux/amd64. Therefore,
+# by leaving it empty we can ensure that the container and binary shipped on it will have the same platform.
+RUN CGO_ENABLED=0 GOOS=${TARGETOS:-linux} GOARCH=${TARGETARCH} go build -a -o nnf-copy-offload daemons/copy-offload/cmd/main.go
+
 ###############################################################################
 FROM builder AS testing
 
+COPY daemons/copy-offload/ daemons/copy-offload/
+
 WORKDIR /workspace
 
 COPY config/ config/
@@ -102,3 +122,24 @@ ENTRYPOINT ["/manager"]
 ARG NNFMFU_TAG_BASE
 ARG NNFMFU_VERSION
 LABEL nnf-mfu="$NNFMFU_TAG_BASE-debug:$NNFMFU_VERSION"
+
+###############################################################################
+FROM $NNFMFU_TAG_BASE:$NNFMFU_VERSION AS copy_offload_production
+
+# The following lines are from the mpiFileUtils (nnf-mfu) Dockerfile;
+# do not change them unless you know what it is you are doing
+RUN sed -i "s/[ #]\(.*StrictHostKeyChecking \).*/ \1no/g" /etc/ssh/ssh_config \
+    && echo "    UserKnownHostsFile /dev/null" >> /etc/ssh/ssh_config
+
+# Copy the executable and execute
+WORKDIR /
+COPY --from=copy_offload_builder /workspace/nnf-copy-offload .
+
+ENTRYPOINT ["/nnf-copy-offload"]
+
+# Make it easy to figure out which nnf-mfu was used.
+#   docker inspect --format='{{json .Config.Labels}}' image:tag
+ARG NNFMFU_TAG_BASE
+ARG NNFMFU_VERSION
+LABEL nnf-mfu="$NNFMFU_TAG_BASE:$NNFMFU_VERSION"
+

Makefile

@@ -47,11 +47,13 @@ BUNDLE_METADATA_OPTS ?= $(BUNDLE_CHANNELS) $(BUNDLE_DEFAULT_CHANNEL)
 # For example, running 'make bundle-build bundle-push catalog-build catalog-push' will build and push both
 # cray.hpe.com/nnf-dm-bundle:$VERSION and cray.hpe.com/nnf-dm-catalog:$VERSION.
 IMAGE_TAG_BASE ?= ghcr.io/nearnodeflash/nnf-dm
+IMAGE_COPY_OFFLOAD_TAG_BASE = $(IMAGE_TAG_BASE)-copy-offload
 IMAGE_TARGET ?= production
+IMAGE_COPY_OFFLOAD_TARGET = copy_offload_$(IMAGE_TARGET)
 
 # The NNF-MFU container image to use in NNFContainerProfile resources.
 NNFMFU_TAG_BASE ?= ghcr.io/nearnodeflash/nnf-mfu
-NNFMFU_VERSION ?= 0.1.3
+NNFMFU_VERSION ?= 0.1.4
 
 CONTAINER_BUILDARGS=--build-arg NNFMFU_TAG_BASE=$(NNFMFU_TAG_BASE) --build-arg NNFMFU_VERSION=$(NNFMFU_VERSION)
 
@@ -62,6 +64,16 @@ BUNDLE_IMG ?= $(IMAGE_TAG_BASE)-bundle:v$(VERSION)
 # ENVTEST_K8S_VERSION refers to the version of kubebuilder assets to be downloaded by envtest binary.
 ENVTEST_K8S_VERSION = 1.29.0
 
+# Tell Kustomize to deploy the default config, or an overlay.
+# To use the 'craystack' overlay:
+#   export KUBECONFIG=/my/craystack/kubeconfig.file
+#   make deploy OVERLAY=craystack
+#
+# To use the 'dp0' overlay:
+#   export KUBECONFIG=/my/dp0/kubeconfig.file
+#   make deploy OVERLAY=dp0
+OVERLAY ?= kind
+
 # Get the currently used golang install path (in GOPATH/bin, unless GOBIN is set)
 ifeq (,$(shell go env GOBIN))
 GOBIN=$(shell go env GOPATH)/bin
@@ -126,7 +138,7 @@ container-unit-test: .version ## Run tests inside a container image
 	${CONTAINER_TOOL} build -f Dockerfile --label $(IMAGE_TAG_BASE)-$@:$(VERSION)-$@ -t $(IMAGE_TAG_BASE)-$@:$(VERSION) --target testing $(CONTAINER_BUILDARGS) .
 	${CONTAINER_TOOL} run --rm -t --name $@-nnf-dm  $(IMAGE_TAG_BASE)-$@:$(VERSION)
 
-##@ Build
+##@ Build the controller manager.
 RPM_PLATFORM ?= linux/amd64
 RPM_TARGET ?= x86_64
 .PHONY: build-daemon-rpm
@@ -158,6 +170,35 @@ build: generate fmt vet ## Build manager binary.
 run: manifests generate fmt vet ## Run a controller from your host.
 	CGO_ENABLED=0 go run cmd/main.go
 
+##@ Build the copy-offload container's daemon outside the container.
+.PHONY: build-copy-offload-local
+build-copy-offload-local: GOOS = $(shell go env GOOS)
+build-copy-offload-local: GOARCH = $(shell go env GOARCH)
+build-copy-offload-local: build-copy-offload-with
+
+.PHONY: build-copy-offload-with
+build-copy-offload-with: $(LOCALBIN)
+build-copy-offload-with: fmt vet ## Build standalone copy-offload daemon
+	CGO_ENABLED=0 GOOS=$(GOOS) GOARCH=$(GOARCH) go build -o bin/nnf-copy-offload daemons/copy-offload/cmd/main.go
+
+.PHONY: build-copy-offload-docker-local
+build-copy-offload-docker-local: GOARCH = $(shell go env GOARCH)
+build-copy-offload-docker-local: build-copy-offload-docker-with
+
+.PHONY: build-copy-offload-docker-amd64
+build-copy-offload-docker-amd64: GOARCH = amd64
+build-copy-offload-docker-amd64: build-copy-offload-docker-with
+
+.PHONY: build-copy-offload-docker-with
+build-copy-offload-docker-with: VERSION ?= $(shell cat .version)
+build-copy-offload-docker-with: .version ## Build docker image with the manager.
+	${CONTAINER_TOOL} build --platform linux/$(GOARCH) --target $(IMAGE_COPY_OFFLOAD_TARGET) -t $(IMAGE_COPY_OFFLOAD_TAG_BASE):$(VERSION) $(CONTAINER_BUILDARGS) .
+
+.PHONY: kind-push-copy-offload
+kind-push-copy-offload: VERSION ?= $(shell cat .version)
+kind-push-copy-offload: .version
+	kind load docker-image $(IMAGE_COPY_OFFLOAD_TAG_BASE):$(VERSION)
+
 # If you wish to build the manager image targeting other platforms you can use the --platform flag.
 # (i.e. docker build --platform linux/arm64). However, you must enable docker buildKit for it.
 # More info: https://docs.docker.com/develop/develop-images/build_enhancements/
@@ -205,12 +246,7 @@ docker-buildx-debug: docker-buildx
 
 kind-push: VERSION ?= $(shell cat .version)
 kind-push: .version ## Push docker image to kind
-	# Nnf-dm is used on all nodes. It's on the management node for the
-	# nnf-dm-controller-manager deployment, and on the rabbit nodes for
-	# the nnf-dm-rsyncnode daemonset that is created by that deployment.
 	kind load docker-image $(IMAGE_TAG_BASE):$(VERSION)
-	${CONTAINER_TOOL} pull gcr.io/kubebuilder/kube-rbac-proxy:v0.13.0
-	kind load docker-image gcr.io/kubebuilder/kube-rbac-proxy:v0.13.0
 
 kind-push-debug: VERSION ?= $(shell cat .version)
 kind-push-debug: IMAGE_TAG_BASE := $(IMAGE_TAG_BASE)-debug
@@ -224,13 +260,13 @@ minikube-push: .version
 
 edit-image: VERSION ?= $(shell cat .version)
 edit-image: .version
-	$(KUSTOMIZE_IMAGE_TAG) config/begin default $(IMAGE_TAG_BASE) $(VERSION) $(NNFMFU_TAG_BASE) $(NNFMFU_VERSION)
+	$(KUSTOMIZE_IMAGE_TAG) config/begin $(OVERLAY) $(IMAGE_TAG_BASE) $(VERSION) $(NNFMFU_TAG_BASE) $(NNFMFU_VERSION)
 
 deploy: kustomize edit-image ## Deploy controller to the K8s cluster specified in ~/.kube/config.
 	./deploy.sh deploy $(KUSTOMIZE) config/begin
 
 undeploy: kustomize ## Undeploy controller from the K8s cluster specified in ~/.kube/config.
-	./deploy.sh undeploy $(KUSTOMIZE) config/default
+	./deploy.sh undeploy $(KUSTOMIZE) config/$(OVERLAY)
 
 # Let .version be phony so that a git update to the workarea can be reflected
 # in it each time it's needed.

cmd/main.go

@@ -42,7 +42,7 @@ import (
 	metricsserver "sigs.k8s.io/controller-runtime/pkg/metrics/server"
 
 	lusv1beta1 "github.com/NearNodeFlash/lustre-fs-operator/api/v1beta1"
-	nnfv1alpha3 "github.com/NearNodeFlash/nnf-sos/api/v1alpha3"
+	nnfv1alpha4 "github.com/NearNodeFlash/nnf-sos/api/v1alpha4"
 
 	controllers "github.com/NearNodeFlash/nnf-dm/internal/controller"
 	//+kubebuilder:scaffold:imports
@@ -57,7 +57,7 @@ func init() {
 	utilruntime.Must(clientgoscheme.AddToScheme(scheme))
 
 	utilruntime.Must(lusv1beta1.AddToScheme(scheme))
-	utilruntime.Must(nnfv1alpha3.AddToScheme(scheme))
+	utilruntime.Must(nnfv1alpha4.AddToScheme(scheme))
 
 	//+kubebuilder:scaffold:scheme
 }
@@ -177,7 +177,7 @@ func (*managerController) GetType() string { return ManagerController }
 func (*managerController) SetOptions(opts *ctrl.Options) {
 	namespaceCache := make(map[string]cache.Config)
 	namespaceCache[corev1.NamespaceDefault] = cache.Config{}
-	namespaceCache[nnfv1alpha3.DataMovementNamespace] = cache.Config{}
+	namespaceCache[nnfv1alpha4.DataMovementNamespace] = cache.Config{}
 	opts.Cache = cache.Options{DefaultNamespaces: namespaceCache}
 }
 
@@ -200,16 +200,16 @@ func (*defaultController) GetType() string { return DefaultController }
 func (*defaultController) SetOptions(opts *ctrl.Options) {
 	namespaceCache := make(map[string]cache.Config)
 	namespaceCache[corev1.NamespaceDefault] = cache.Config{}
-	namespaceCache[nnfv1alpha3.DataMovementNamespace] = cache.Config{}
-	namespaceCache[nnfv1alpha3.DataMovementProfileNamespace] = cache.Config{}
+	namespaceCache[nnfv1alpha4.DataMovementNamespace] = cache.Config{}
+	namespaceCache[nnfv1alpha4.DataMovementProfileNamespace] = cache.Config{}
 	opts.Cache = cache.Options{DefaultNamespaces: namespaceCache}
 }
 
 func (c *defaultController) SetupReconcilers(mgr manager.Manager) (err error) {
 	if err = (&controllers.DataMovementReconciler{
 		Client:         mgr.GetClient(),
 		Scheme:         mgr.GetScheme(),
-		WatchNamespace: nnfv1alpha3.DataMovementNamespace,
+		WatchNamespace: nnfv1alpha4.DataMovementNamespace,
 	}).SetupWithManager(mgr); err != nil {
 		setupLog.Error(err, "unable to create controller", "controller", c.GetType())
 		os.Exit(1)
@@ -227,8 +227,8 @@ func (*nodeController) GetType() string { return NodeController }
 func (*nodeController) SetOptions(opts *ctrl.Options) {
 	namespaceCache := make(map[string]cache.Config)
 	namespaceCache[corev1.NamespaceDefault] = cache.Config{}
-	namespaceCache[nnfv1alpha3.DataMovementNamespace] = cache.Config{}
-	namespaceCache[nnfv1alpha3.DataMovementProfileNamespace] = cache.Config{}
+	namespaceCache[nnfv1alpha4.DataMovementNamespace] = cache.Config{}
+	namespaceCache[nnfv1alpha4.DataMovementProfileNamespace] = cache.Config{}
 	namespaceCache[os.Getenv("NNF_NODE_NAME")] = cache.Config{}
 	opts.Cache = cache.Options{DefaultNamespaces: namespaceCache}
 }

config/copy-offload/copy_offload_role.yaml

@@ -0,0 +1,62 @@
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: copy-offload-role
+rules:
+- apiGroups:
+  - lus.cray.hpe.com
+  resources:
+  - lustrefilesystems
+  verbs:
+  - get
+  - list
+  - watch
+- apiGroups:
+  - dataworkflowservices.github.io
+  resources:
+  - workflows
+  verbs:
+  - get
+  - list
+  - watch
+- apiGroups:
+  - dataworkflowservices.github.io
+  resources:
+  - clientmounts
+  verbs:
+  - get
+  - list
+  - watch
+- apiGroups:
+  - nnf.cray.hpe.com
+  resources:
+  - nnfdatamovementprofiles
+  verbs:
+  - get
+  - list
+  - watch
+- apiGroups:
+  - nnf.cray.hpe.com
+  resources:
+  - nnfnodes
+  verbs:
+  - get
+  - list
+  - watch
+- apiGroups:
+  - nnf.cray.hpe.com
+  resources:
+  - nnfnodestorages
+  verbs:
+  - get
+  - list
+  - watch
+- apiGroups:
+  - nnf.cray.hpe.com
+  resources:
+  - nnfstorages
+  verbs:
+  - get
+  - list
+  - watch

config/copy-offload/copy_offload_role_binding.yaml

@@ -0,0 +1,12 @@
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: copy-offload-rolebinding
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: copy-offload-role
+subjects:
+- kind: ServiceAccount
+  name: copy-offload
+  namespace: default

config/copy-offload/copy_offload_service_account.yaml

@@ -0,0 +1,18 @@
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: copy-offload
+  namespace: default
+---
+# As of Kubernetes 1.24, ServiceAccount tokens are no longer automatically
+# generated. Instead, manually create the secret and the token key in the
+# data field will be automatically set.
+apiVersion: v1
+kind: Secret
+metadata:
+  name: copy-offload
+  namespace: default
+  annotations:
+    kubernetes.io/service-account.name: copy-offload
+    kubernetes.io/service-account.namespace: default
+type: kubernetes.io/service-account-token

config/copy-offload/kustomization.yaml

@@ -0,0 +1,14 @@
+# Value of this field is prepended to the
+# names of all resources, e.g. a deployment named
+# "wordpress" becomes "alices-wordpress".
+# Note that it should also match with the prefix (text before '-') of the namespace
+# field above.
+namePrefix: nnf-dm-
+
+resources:
+- copy_offload_service_account.yaml
+- copy_offload_role.yaml
+- copy_offload_role_binding.yaml
+
+configurations:
+  - kustomizeconfig.yaml

config/copy-offload/kustomizeconfig.yaml

@@ -0,0 +1,4 @@
+namePrefix:
+- path: metadata/annotations/kubernetes.io\/service-account.name
+namespace:
+- path: metadata/annotations/kubernetes.io\/service-account.namespace

config/dp0/kustomization.yaml

@@ -0,0 +1,6 @@
+resources:
+- ../top
+
+patches:
+# Arguments for the controller manager that are specific to dp0
+- path: manager_environment_deploy_patch.yaml