Feat makefile help (sustainable-computing-io#1332)

* Makefile: Add usage and more image build options This change introduces a helpful message for a user that issues a `make help` command. It also adds options to build the container images with/without dcgm explicitly. This will help improve build time for developers that aren't using dcgm. Signed-off-by: Maryam Tahhan <[email protected]> * Dockerfile: remove trailing whitespace Signed-off-by: Maryam Tahhan <[email protected]> * Docker: fix build-arg INSTALL_DCGM=false The dockerfile was only checking that the INSTALL_DCGM build argument was set and not it's value. So if INSTALL_DCGM=false was being passed, it considered the build argument as set. This change explicitly checks the value as true. Signed-off-by: Maryam Tahhan <[email protected]> --------- Signed-off-by: Maryam Tahhan <[email protected]>
maryamtahhan · Apr 8, 2024 · 3d1d7ab · 3d1d7ab
1 parent c6080eb
commit 3d1d7ab
Show file tree

Hide file tree

Showing 2 changed files with 61 additions and 40 deletions.
diff --git a/Makefile b/Makefile
@@ -2,6 +2,24 @@ all: kepler
 
 include bpfassets/libbpf/Makefile
 
+##@ Help
+
+# The help target prints out all targets with their descriptions organized
+# beneath their categories. The categories are represented by '##@' and the
+# target descriptions by '##'. The awk commands is responsible for reading the
+# entire set of makefiles included in this invocation, looking for lines of the
+# file as xyz: ## something, and then pretty-format the target and help. Then,
+# if there's a line with ##@ something, that gets pretty-printed as a category.
+# More info on the usage of ANSI control characters for terminal formatting:
+# https://en.wikipedia.org/wiki/ANSI_escape_code#SGR_parameters
+# More info on the awk command:
+# http://linuxcommand.org/lc3_adv_awk.php
+
+.PHONY: help
+help: ## Display this help.
+	@awk 'BEGIN {FS = ":.*##"; printf "\nUsage:\n  make \033[36m<target>\033[0m\n"} /^[a-zA-Z_0-9-]+:.*?##/ { printf "  \033[36m%-15s\033[0m %s\n", $$1, $$2 } /^##@/ { printf "\n\033[1m%s\033[0m\n", substr($$0, 5) } ' $(MAKEFILE_LIST)
+
+
 ### env define ###
 export BIN_TIMESTAMP ?=$(shell date -u +'%Y-%m-%dT%H:%M:%SZ')
 export TIMESTAMP ?=$(shell echo $(BIN_TIMESTAMP) | tr -d ':' | tr 'T' '-' | tr -d 'Z')
@@ -68,83 +86,85 @@ ifndef GOBIN
 	GOBIN := $(GOPATH)/bin
 endif
 
-# NOTE: project related tools get installed to tmp dir which is ignored by 
+# NOTE: project related tools get installed to tmp dir which is ignored by
 PROJECT_DIR := $(shell dirname $(abspath $(firstword $(MAKEFILE_LIST))))
 TOOLS_DIR=$(PROJECT_DIR)/tmp/bin
 KUSTOMIZE = $(TOOLS_DIR)/kustomize
 GOVULNCHECK = $(TOOLS_DIR)/govulncheck
 
 base_dir := $(patsubst %/,%,$(dir $(realpath $(firstword $(MAKEFILE_LIST)))))
 
-### Default ###
-kepler: build_containerized
+##@ Default
+kepler: build_containerized ## Build Kepler.
 .PHONY: kepler
 
 clean: clean-cross-build
 .PHONY: clean
 
-### build container ###
-build_containerized: tidy-vendor format
+##@ Container build.
+image_builder_check:
 	@if [ -z '$(CTR_CMD)' ] ; then echo '!! ERROR: containerized builds require podman||docker CLI, none found $$PATH' >&2 && exit 1; fi
-	echo BIN_TIMESTAMP==$(BIN_TIMESTAMP)
+
+build_image: image_builder_check ## Build image without DCGM.
 	# build kepler without dcgm
 	$(CTR_CMD) build -t $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_BUILD_TAG) \
 		-f $(DOCKERFILE) \
 		--network host \
-		--build-arg BIN_TIMESTAMP=$(BIN_TIMESTAMP) \
+		--build-arg INSTALL_DCGM="false" \
 		--build-arg VERSION=$(VERSION) \
 		--platform="linux/$(GOARCH)" \
 		.
-
 	$(CTR_CMD) tag $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_BUILD_TAG) $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_TAG)
+.PHONY: build_image
 
+build_image_dcgm:  image_builder_check ## Build image with DCGM.
 	# build kepler with dcgm
 	$(CTR_CMD) build -t $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_BUILD_TAG)-"dcgm" \
 		-f $(DOCKERFILE) \
 		--network host \
-		--build-arg BIN_TIMESTAMP=$(BIN_TIMESTAMP) \
 		--build-arg INSTALL_DCGM="true" \
 		--build-arg VERSION=$(VERSION) \
 		--platform="linux/$(GOARCH)" \
 		.
-
 	$(CTR_CMD) tag $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_BUILD_TAG)-dcgm $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_TAG)-dcgm
+.PHONY: build_image_dcgm
 
+build_containerized: tidy-vendor format build_image build_image_dcgm  ## Build ALL container images.
 .PHONY: build_containerized
 
-save-image:
+save-image: ## Save container image.
 	@mkdir -p _output
 	$(CTR_CMD) save $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_TAG) | gzip > "${IMAGE_OUTPUT_PATH}"
 .PHONY: save-image
 
-load-image:
+load-image: ## Load container image.
 	$(CTR_CMD) load -i "${INPUT_PATH}"
 .PHONY: load-image
 
-image-prune:
+image-prune: ## Image prune.
 	$(CTR_CMD) image prune -a -f || true
 .PHONY: image-prune
 
-push-image:
+push-image:  ## Push image.
 	$(CTR_CMD) push $(CTR_CMD_PUSH_OPTIONS) $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_TAG)
 .PHONY: push-image
 
+##@ General build
 clean-cross-build:
 	$(RM) -r '$(CROSS_BUILD_BINDIR)'
 	$(RM) -rf $(OUTPUT_DIR)/staging
 	if [ -d '$(OUTPUT_DIR)' ]; then rmdir --ignore-fail-on-non-empty '$(OUTPUT_DIR)'; fi
 .PHONY: clean-cross-build
 
-### build binary ###
-build: clean_build_local _build_local copy_build_local
+build: clean_build_local _build_local copy_build_local ##  Build binary and copy to $(OUTPUT_DIR)/bin
 .PHONY: build
 
-_build_local: genlibbpf
+_build_local: genlibbpf ##  Build Kepler binary locally.
 	@echo TAGS=$(GO_BUILD_TAGS)
 	@mkdir -p "$(CROSS_BUILD_BINDIR)/$(GOOS)_$(GOARCH)"
 	+@$(GOENV) go build -v -tags ${GO_BUILD_TAGS} -o $(CROSS_BUILD_BINDIR)/$(GOOS)_$(GOARCH)/kepler -ldflags $(LDFLAGS) ./cmd/exporter/exporter.go
 
-container_build:
+container_build: ## Run a container and build Kepler inside it.
 	$(CTR_CMD) run --rm \
 		-v $(base_dir):/kepler:Z -w /kepler \
 		--user $(shell id -u):$(shell id -g) \
@@ -153,13 +173,14 @@ container_build:
 		$(BUILDER_IMAGE) \
 		git config --add safe.directory /kepler && make build
 
-build_rpm:
+##@ RPM build
+build_rpm: ## Build the Kepler Binary RPM.
 	rpmbuild packaging/rpm/kepler.spec --build-in-place -bb
 
-build_container_rpm:
+build_container_rpm: ## Build the Containerized Kepler RPM.
 	rpmbuild packaging/rpm/container-kepler.spec --build-in-place -bb
 
-containerized_build_rpm:
+containerized_build_rpm: ## Build the Kepler Binary RPM inside a container.
 	@mkdir -p $(base_dir)/$(OUTPUT_DIR)/rpmbuild
 	$(CTR_CMD) run --rm \
 		-v $(base_dir):/kepler:Z -w /kepler -v $(base_dir)/$(OUTPUT_DIR)/rpmbuild:/opt/app-root/src/rpmbuild \
@@ -169,7 +190,7 @@ containerized_build_rpm:
 		$(BUILDER_IMAGE) \
 		make build_rpm
 
-containerized_build_container_rpm:
+containerized_build_container_rpm: ## Build the Containerized Kepler RPM inside a container.
 	@mkdir -p $(base_dir)/$(OUTPUT_DIR)/rpmbuild
 	$(CTR_CMD) run --rm \
 		-v $(base_dir):/kepler:Z -w /kepler -v $(base_dir)/$(OUTPUT_DIR)/rpmbuild:/opt/app-root/src/rpmbuild \
@@ -195,10 +216,10 @@ cross-build-linux-s390x:
 	+$(MAKE) _build_local GOOS=linux GOARCH=s390x
 .PHONY: cross-build-linux-s390x
 
-cross-build: clean_build_local cross-build-linux-amd64 cross-build-linux-arm64 cross-build-linux-s390x copy_build_local
+cross-build: clean_build_local cross-build-linux-amd64 cross-build-linux-arm64 cross-build-linux-s390x copy_build_local ## Build Kepler for multiple archs.
 .PHONY: cross-build
 
-### toolkit ###
+## toolkit ###
 tidy-vendor:
 	go mod tidy -v
 	go mod vendor
@@ -246,7 +267,7 @@ test-container-verbose: ginkgo-set tidy-vendor
 		-covermode=atomic -coverprofile=coverage.out \
 		-v $$(go list ./... | grep pkg | grep -v bpfassets) \
 		--race -cover --count=1 --vet=all
-	
+
 test-mac-verbose: ginkgo-set
 	@echo TAGS=$(GO_TEST_TAGS)
 	@go test $$(go list ./... | grep pkg | grep -v bpfassets) --race --bench=. -cover --count=1 --vet=all
@@ -288,56 +309,56 @@ build-manifest: kustomize
 	./hack/build-manifest.sh "${OPTS}"
 .PHONY: build-manifest
 
-##@ Development env
+##@ Development Env
 CLUSTER_PROVIDER ?= kind
 LOCAL_DEV_CLUSTER_VERSION ?= main
 KIND_WORKER_NODES ?=2
 
-cluster-clean: build-manifest
+cluster-clean: build-manifest ## Undeploy Kepler in the cluster.
 	./hack/cluster-clean.sh
 .PHONY: cluster-clean
 
-cluster-deploy:
+cluster-deploy: ## Deploy Kepler in the cluster.
 	./hack/cluster-deploy.sh
 .PHONY: cluster-deploy
 
-cluster-up:
+cluster-up:  ## Create the Kind cluster
 	CLUSTER_PROVIDER=$(CLUSTER_PROVIDER) \
 	LOCAL_DEV_CLUSTER_VERSION=$(LOCAL_DEV_CLUSTER_VERSION) \
 	KIND_WORKER_NODES=$(KIND_WORKER_NODES) \
 	./hack/cluster.sh up
 .PHONY: cluster-up
 
-cluster-down:
+cluster-down: ## Delete the Kind cluster.
 	CLUSTER_PROVIDER=$(CLUSTER_PROVIDER) \
 	LOCAL_DEV_CLUSTER_VERSION=$(LOCAL_DEV_CLUSTER_VERSION) \
 	KIND_WORKER_NODES=$(KIND_WORKER_NODES) \
 	./hack/cluster.sh down
 .PHONY: cluster-down
 
-cluster-restart:
+cluster-restart: ## Restart the Kind cluster.
 	CLUSTER_PROVIDER=$(CLUSTER_PROVIDER) \
 	LOCAL_DEV_CLUSTER_VERSION=$(LOCAL_DEV_CLUSTER_VERSION) \
 	KIND_WORKER_NODES=$(KIND_WORKER_NODES) \
 	./hack/cluster.sh restart
 .PHONY: cluster-restart
 
-e2e:
+e2e: # Run E2E integration tests.
 	./hack/verify.sh integration
 .PHONY: e2e
 
-### platform-validation ###
+##@ platform-validation
 
 VALIDATION_DOCKERFILE := $(SRC_ROOT)/build/Dockerfile.kepler-validator
 
-build-validator: tidy-vendor format
+build-validator: tidy-vendor format ## Build Validator.
 	@echo TAGS=$(GO_BUILD_TAGS)
 	@mkdir -p "$(CROSS_BUILD_BINDIR)/$(GOOS)_$(GOARCH)"
 	+@$(GOENV) go build -v -tags ${GO_BUILD_TAGS} -o $(CROSS_BUILD_BINDIR)/$(GOOS)_$(GOARCH)/validator -ldflags $(LDFLAGS) ./cmd/validator/validator.go
 	cp $(CROSS_BUILD_BINDIR)/$(GOOS)_$(GOARCH)/validator $(CROSS_BUILD_BINDIR)
 .PHONY: build-validator
 
-build-validation-container:
+build-validation-container: ## Build validation Container.
 	$(CTR_CMD) build -t $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_TAG) \
 		-f $(VALIDATION_DOCKERFILE) .
 .PHONY: build-validation-container
@@ -350,7 +371,7 @@ get-env:
 	$(CTR_CMD) run -i --rm -v $(SRC_ROOT)/e2e/platform-validation:/output $(IMAGE_REPO)/$(IMAGE_NAME):$(IMAGE_TAG) /usr/bin/validator -gen-env=true
 .PHONY: get-env
 
-platform-validation: ginkgo-set get-env
+platform-validation: ginkgo-set get-env ## Run Kepler platform validation.
 	./hack/verify.sh platform
 .PHONY: platform-validation
 

diff --git a/build/Dockerfile b/build/Dockerfile
@@ -15,18 +15,18 @@ RUN yum -y update
 
 ENV NVIDIA_VISIBLE_DEVICES=all
 ENV NVIDIA_DRIVER_CAPABILITIES=utility
-ENV NVIDIA_MIG_CONFIG_DEVICES=all 
-ENV NVIDIA_MIG_MONITOR_DEVICES=all 
+ENV NVIDIA_MIG_CONFIG_DEVICES=all
+ENV NVIDIA_MIG_MONITOR_DEVICES=all
 
 RUN INSTALL_PKGS=" \
     libbpf \
     " && \
-	yum install -y $INSTALL_PKGS 
+	yum install -y $INSTALL_PKGS
 
 RUN if [ "$TARGETARCH" == "amd64" ]; then \
         yum install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm; \
         yum install -y cpuid; \
-        if [[ ! -z "$INSTALL_DCGM" ]]; then \
+        if [[ ! -z "$INSTALL_DCGM" && "$INSTALL_DCGM" == "true" ]]; then \
             dnf config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/rhel9/x86_64/cuda-rhel9.repo; \
             yum install -y datacenter-gpu-manager; \
         fi; \