feature(controller): remove dockerd executable (#432)

Prevents arguments from being split when e.g. the RUNNER_GROUP variable contains spaces (which is legit. One can create such groups in GitHub).

I've seen that all workers with group names that contain no spaces can register successfully, while all workers with groups that contain spaces will not register.

Furthermore, I suppose also other chars can be used here to inject arbitrary commands in an unsupported way via e.g. pipe symbol.

Quoting the vars correctly should prevent that and allow for e.g. group names and runner labels with spaces and other bash reserved characters.

Makefile

@@ -14,6 +14,8 @@ else
 GOBIN=$(shell go env GOBIN)
 endif
 
+TEST_ASSETS=$(PWD)/test-assets
+
 # default list of platforms for which multiarch image is built
 ifeq (${PLATFORMS}, )
 	export PLATFORMS="linux/amd64,linux/arm64"
@@ -37,6 +39,13 @@ all: manager
 test: generate fmt vet manifests
 	go test ./... -coverprofile cover.out
 
+test-with-deps: kube-apiserver etcd kubectl
+	# See https://pkg.go.dev/sigs.k8s.io/controller-runtime/pkg/envtest#pkg-constants
+	TEST_ASSET_KUBE_APISERVER=$(KUBE_APISERVER_BIN) \
+	TEST_ASSET_ETCD=$(ETCD_BIN) \
+	TEST_ASSET_KUBECTL=$(KUBECTL_BIN) \
+	  make test
+
 # Build manager binary
 manager: generate fmt vet
 	go build -o bin/manager main.go
@@ -191,3 +200,66 @@ ifeq (, $(wildcard $(GOBIN)/yq))
 	}
 endif
 YQ=$(GOBIN)/yq
+
+OS_NAME := $(shell uname -s | tr A-Z a-z)
+
+# find or download etcd
+etcd:
+ifeq (, $(wildcard $(TEST_ASSETS)/etcd))
+	@{ \
+	set -xe ;\
+	INSTALL_TMP_DIR=$$(mktemp -d) ;\
+	cd $$INSTALL_TMP_DIR ;\
+	wget https://github.com/kubernetes-sigs/kubebuilder/releases/download/v2.3.2/kubebuilder_2.3.2_$(OS_NAME)_amd64.tar.gz ;\
+	mkdir -p $(TEST_ASSETS) ;\
+	tar zxvf kubebuilder_2.3.2_$(OS_NAME)_amd64.tar.gz ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/etcd $(TEST_ASSETS)/etcd ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/kube-apiserver $(TEST_ASSETS)/kube-apiserver ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/kubectl $(TEST_ASSETS)/kubectl ;\
+	rm -rf $$INSTALL_TMP_DIR ;\
+	}
+ETCD_BIN=$(TEST_ASSETS)/etcd
+else
+ETCD_BIN=$(TEST_ASSETS)/etcd
+endif
+
+# find or download kube-apiserver
+kube-apiserver:
+ifeq (, $(wildcard $(TEST_ASSETS)/kube-apiserver))
+	@{ \
+	set -xe ;\
+	INSTALL_TMP_DIR=$$(mktemp -d) ;\
+	cd $$INSTALL_TMP_DIR ;\
+	wget https://github.com/kubernetes-sigs/kubebuilder/releases/download/v2.3.2/kubebuilder_2.3.2_$(OS_NAME)_amd64.tar.gz ;\
+	mkdir -p $(TEST_ASSETS) ;\
+	tar zxvf kubebuilder_2.3.2_$(OS_NAME)_amd64.tar.gz ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/etcd $(TEST_ASSETS)/etcd ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/kube-apiserver $(TEST_ASSETS)/kube-apiserver ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/kubectl $(TEST_ASSETS)/kubectl ;\
+	rm -rf $$INSTALL_TMP_DIR ;\
+	}
+KUBE_APISERVER_BIN=$(TEST_ASSETS)/kube-apiserver
+else
+KUBE_APISERVER_BIN=$(TEST_ASSETS)/kube-apiserver
+endif
+
+
+# find or download kubectl
+kubectl:
+ifeq (, $(wildcard $(TEST_ASSETS)/kubectl))
+	@{ \
+	set -xe ;\
+	INSTALL_TMP_DIR=$$(mktemp -d) ;\
+	cd $$INSTALL_TMP_DIR ;\
+	wget https://github.com/kubernetes-sigs/kubebuilder/releases/download/v2.3.2/kubebuilder_2.3.2_$(OS_NAME)_amd64.tar.gz ;\
+	mkdir -p $(TEST_ASSETS) ;\
+	tar zxvf kubebuilder_2.3.2_$(OS_NAME)_amd64.tar.gz ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/etcd $(TEST_ASSETS)/etcd ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/kube-apiserver $(TEST_ASSETS)/kube-apiserver ;\
+	mv kubebuilder_2.3.2_$(OS_NAME)_amd64/bin/kubectl $(TEST_ASSETS)/kubectl ;\
+	rm -rf $$INSTALL_TMP_DIR ;\
+	}
+KUBECTL_BIN=$(TEST_ASSETS)/kubectl
+else
+KUBECTL_BIN=$(TEST_ASSETS)/kubectl
+endif

README.md

@@ -163,7 +163,7 @@ Log-in to a GitHub account that has `admin` privileges for the repository, and [
 
 * repo (Full control)
 
-**Scopes for a Organisation Runner**
+**Scopes for a Organization Runner**
 
 * repo (Full control)
 * admin:org (Full control)
@@ -419,11 +419,11 @@ spec:
 > Please get prepared to put some time and effort to learn and leverage this feature!
 
 `actions-runner-controller` has an optional Webhook server that receives GitHub Webhook events and scale
-[`RunnerDeployment`s](#runnerdeployments) by updating corresponding [`HorizontalRunnerAutoscaler`s](#autoscaling).
+[`RunnerDeployments`](#runnerdeployments) by updating corresponding [`HorizontalRunnerAutoscalers`](#autoscaling).
 
 Today, the Webhook server can be configured to respond GitHub `check_run`, `pull_request`, and `push` events
 by scaling up the matching `HorizontalRunnerAutoscaler` by N replica(s), where `N` is configurable within
-`HorizontalRunerAutoscaler`'s `Spec`.
+`HorizontalRunerAutoscaler's` `Spec`.
 
 More concretely, you can configure the targeted GitHub event types and the `N` in
 `scaleUpTriggers`:

acceptance/testdata/runnerdeploy.yaml

@@ -7,3 +7,14 @@ spec:
   template:
     spec:
       repository: mumoshu/actions-runner-controller-ci
+      #
+      # dockerd within runner container
+      #
+      ## Replace `mumoshu/actions-runner-dind:dev` with your dind image
+      #dockerdWithinRunnerContainer: true
+      #image: mumoshu/actions-runner-dind:dev
+
+      #
+      # Set the MTU used by dockerd-managed network interfaces (including docker-build)
+      #
+      #dockerMTU: 1450

controllers/runner_controller.go

@@ -634,45 +634,58 @@ func (r *RunnerReconciler) newPod(runner v1alpha1.Runner) (corev1.Pod, error) {
 		}...)
 	}
 
-	if !dockerdInRunner && dockerEnabled {
-		runnerVolumeName := "runner"
-		runnerVolumeMountPath := "/runner"
+	//
+	// /runner must be generated on runtime from /runnertmp embedded in the container image.
+	//
+	// When you're NOT using dindWithinRunner=true,
+	// it must also be shared with the dind container as it seems like required to run docker steps.
+	//
 
-		pod.Spec.Volumes = []corev1.Volume{
-			{
+	runnerVolumeName := "runner"
+	runnerVolumeMountPath := "/runner"
+
+	pod.Spec.Volumes = append(pod.Spec.Volumes,
+		corev1.Volume{
+			Name: runnerVolumeName,
+			VolumeSource: corev1.VolumeSource{
+				EmptyDir: &corev1.EmptyDirVolumeSource{},
+			},
+		},
+	)
+
+	pod.Spec.Containers[0].VolumeMounts = append(pod.Spec.Containers[0].VolumeMounts,
+		corev1.VolumeMount{
+			Name:      runnerVolumeName,
+			MountPath: runnerVolumeMountPath,
+		},
+	)
+
+	if !dockerdInRunner && dockerEnabled {
+		pod.Spec.Volumes = append(pod.Spec.Volumes,
+			corev1.Volume{
 				Name: "work",
 				VolumeSource: corev1.VolumeSource{
 					EmptyDir: &corev1.EmptyDirVolumeSource{},
 				},
 			},
-			{
-				Name: runnerVolumeName,
-				VolumeSource: corev1.VolumeSource{
-					EmptyDir: &corev1.EmptyDirVolumeSource{},
-				},
-			},
-			{
+			corev1.Volume{
 				Name: "certs-client",
 				VolumeSource: corev1.VolumeSource{
 					EmptyDir: &corev1.EmptyDirVolumeSource{},
 				},
 			},
-		}
-		pod.Spec.Containers[0].VolumeMounts = []corev1.VolumeMount{
-			{
+		)
+		pod.Spec.Containers[0].VolumeMounts = append(pod.Spec.Containers[0].VolumeMounts,
+			corev1.VolumeMount{
 				Name:      "work",
 				MountPath: workDir,
 			},
-			{
-				Name:      runnerVolumeName,
-				MountPath: runnerVolumeMountPath,
-			},
-			{
+			corev1.VolumeMount{
 				Name:      "certs-client",
 				MountPath: "/certs/client",
 				ReadOnly:  true,
 			},
-		}
+		)
 		pod.Spec.Containers[0].Env = append(pod.Spec.Containers[0].Env, []corev1.EnvVar{
 			{
 				Name:  "DOCKER_HOST",
@@ -718,11 +731,17 @@ func (r *RunnerReconciler) newPod(runner v1alpha1.Runner) (corev1.Pod, error) {
 
 		if mtu := runner.Spec.DockerMTU; mtu != nil {
 			pod.Spec.Containers[1].Env = append(pod.Spec.Containers[1].Env, []corev1.EnvVar{
+				// See https://docs.docker.com/engine/security/rootless/
 				{
 					Name:  "DOCKERD_ROOTLESS_ROOTLESSKIT_MTU",
 					Value: fmt.Sprintf("%d", *runner.Spec.DockerMTU),
 				},
 			}...)
+
+			pod.Spec.Containers[1].Args = append(pod.Spec.Containers[1].Args,
+				"--mtu",
+				fmt.Sprintf("%d", *runner.Spec.DockerMTU),
+			)
 		}
 
 	}

runner/entrypoint.sh

@@ -29,21 +29,13 @@ else
   exit 1
 fi
 
-if [ -n "${RUNNER_WORKDIR}" ]; then
-  WORKDIR_ARG="--work ${RUNNER_WORKDIR}"
-fi
-
-if [ -n "${RUNNER_LABELS}" ]; then
-  LABEL_ARG="--labels ${RUNNER_LABELS}"
-fi
-
 if [ -z "${RUNNER_TOKEN}" ]; then
   echo "RUNNER_TOKEN must be set" 1>&2
   exit 1
 fi
 
 if [ -z "${RUNNER_REPO}" ] && [ -n "${RUNNER_GROUP}" ];then
-  RUNNER_GROUP_ARG="--runnergroup ${RUNNER_GROUP}"
+  RUNNER_GROUPS=${RUNNER_GROUP}
 fi
 
 # Hack due to https://github.com/summerwind/actions-runner-controller/issues/252#issuecomment-758338483
@@ -56,7 +48,14 @@ sudo chown -R runner:docker /runner
 mv /runnertmp/* /runner/
 
 cd /runner
-./config.sh --unattended --replace --name "${RUNNER_NAME}" --url "${GITHUB_URL}${ATTACH}" --token "${RUNNER_TOKEN}" ${RUNNER_GROUP_ARG} ${LABEL_ARG} ${WORKDIR_ARG}
+./config.sh --unattended --replace \
+  --name "${RUNNER_NAME}" \
+  --url "${GITHUB_URL}${ATTACH}" \
+  --token "${RUNNER_TOKEN}" \
+  --runnergroup "${RUNNER_GROUPS}" \
+  --labels "${RUNNER_LABELS}" \
+  --work "${RUNNER_WORKDIR}"
+
 mkdir ./externals
 # Hack due to the DinD volumes
 mv ./externalstmp/* ./externals/

runner/startup.sh

@@ -17,6 +17,34 @@ function wait_for_process () {
     return 0
 }
 
+sudo /bin/bash <<SCRIPT
+mkdir -p /etc/docker
+
+cat <<EOS > /etc/docker/daemon.json
+{
+EOS
+
+if [ -n "${MTU}" ]; then
+cat <<EOS >> /etc/docker/daemon.json
+  "mtu": ${MTU}
+EOS
+# See https://docs.docker.com/engine/security/rootless/
+echo "environment=DOCKERD_ROOTLESS_ROOTLESSKIT_MTU=${MTU}" >> /etc/supervisor/conf.d/dockerd.conf
+fi
+
+cat <<EOS >> /etc/docker/daemon.json
+}
+EOS
+SCRIPT
+
+INFO "Using /etc/docker/daemon.json with the following content"
+
+cat /etc/docker/daemon.json
+
+INFO "Using /etc/supervisor/conf.d/dockerd.conf with the following content"
+
+cat /etc/supervisor/conf.d/dockerd.conf
+
 INFO "Starting supervisor"
 sudo /usr/bin/supervisord -n >> /dev/null 2>&1 &
 
@@ -27,6 +55,8 @@ for process in "${processes[@]}"; do
     wait_for_process "$process"
     if [ $? -ne 0 ]; then
         ERROR "$process is not running after max time"
+        ERROR "Dumping /var/log/dockerd.err.log to help investigation"
+        cat /var/log/dockerd.err.log
         exit 1
     else 
         INFO "$process is running"