feat: improve CSI robustness

.gitlab/CODEOWNERS

@@ -4,9 +4,9 @@
 .gitlab/CODEOWNERS @lanwall12 @nitper @bsherrycrusoe @ksosnowski3 @dpattishall
 
 # more restrictive ownership for adding dependencies and changing ci/build
-.gitlab-ci.yml     @lanwall12 @nitper @crusoeenergy/island/ccx
-.golangci-lint.yml @lanwall12 @nitper @crusoeenergy/island/ccx
-go.mod             @lanwall12 @nitper @crusoeenergy/island/ccx
-Makefile           @lanwall12 @nitper @crusoeenergy/island/ccx
-Dockerfile         @lanwall12 @nitper @crusoeenergy/island/ccx
-Dockerfile.*       @lanwall12 @nitper @crusoeenergy/island/ccx
+.gitlab-ci.yml     @lanwall12 @nitper @crusoeenergy/island/managed
+.golangci-lint.yml @lanwall12 @nitper @crusoeenergy/island/managed
+go.mod             @lanwall12 @nitper @crusoeenergy/island/managed
+Makefile           @lanwall12 @nitper @crusoeenergy/island/managed
+Dockerfile         @lanwall12 @nitper @crusoeenergy/island/managed
+Dockerfile.*       @lanwall12 @nitper @crusoeenergy/island/managed

.golangci.yml

@@ -103,7 +103,7 @@ issues:
       source: "^//go:generate "
 
 run:
-  # include test  files or not
+  # include test files or not
   tests: true
 
 # golangci.com configuration

go.mod

@@ -6,6 +6,7 @@ require (
 	github.com/antihax/optional v1.0.0
 	github.com/container-storage-interface/spec v1.10.0
 	github.com/crusoecloud/client-go v0.1.59
+	github.com/google/uuid v1.6.0
 	github.com/spf13/cobra v1.8.1
 	github.com/spf13/viper v1.19.0
 	github.com/thediveo/enumflag/v2 v2.0.5
@@ -34,7 +35,6 @@ require (
 	github.com/google/gnostic-models v0.6.8 // indirect
 	github.com/google/go-cmp v0.6.0 // indirect
 	github.com/google/gofuzz v1.2.0 // indirect
-	github.com/google/uuid v1.6.0 // indirect
 	github.com/hashicorp/hcl v1.0.0 // indirect
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
 	github.com/josharian/intern v1.0.0 // indirect

internal/controller/util.go

@@ -22,6 +22,7 @@ var (
 		csi.VolumeCapability_AccessMode_SINGLE_NODE_READER_ONLY:   {},
 		csi.VolumeCapability_AccessMode_SINGLE_NODE_SINGLE_WRITER: {},
 		csi.VolumeCapability_AccessMode_SINGLE_NODE_MULTI_WRITER:  {},
+		csi.VolumeCapability_AccessMode_MULTI_NODE_READER_ONLY:    {},
 	}
 	//nolint:gochecknoglobals  // can't construct const map
 	fsAllowedAccessModes = map[csi.VolumeCapability_AccessMode_Mode]struct{}{

internal/crusoe/auth.go

@@ -74,7 +74,7 @@ func signMessageV1_0(message []byte, encodedKey string) ([]byte, error) {
 	// Key is b64 encoded.
 	expectedKey, err := base64.RawURLEncoding.DecodeString(encodedKey)
 	if err != nil {
-		return nil, fmt.Errorf("failed to decode key: %w", err)
+		return nil, fmt.Errorf("failed to decode Crusoe Secret Key: %w", err)
 	}
 
 	mac := hmac.New(sha256.New, expectedKey)

internal/server.go

@@ -13,6 +13,8 @@ import (
 	"sync"
 	"syscall"
 
+	"github.com/google/uuid"
+
 	"github.com/crusoecloud/crusoe-csi-driver/internal/controller"
 
 	"github.com/crusoecloud/crusoe-csi-driver/internal/common"
@@ -42,15 +44,16 @@ const (
 	projectIDEnvKey   = "CRUSOE_PROJECT_ID"
 	projectIDLabelKey = "crusoe.ai/project.id"
 
-	numExpectedComponents = 2
+	vmIDFilePath = "/sys/class/dmi/id/product_uuid"
 )
 
 var (
 	errInstanceNotFound  = errors.New("instance not found")
 	errMultipleInstances = errors.New("multiple instances found")
+	errVMIDReadFailed    = fmt.Errorf("failed to read %s for VM ID", vmIDFilePath)
+	errVMIDParseFailed   = fmt.Errorf("failed to parse %s for VM ID", vmIDFilePath)
 	errProjectIDNotFound = fmt.Errorf("project ID not found in %s env var or %s node label",
 		projectIDEnvKey, projectIDLabelKey)
-	errInvalidNodeName = errors.New("invalid node name")
 )
 
 func interruptHandler() (*sync.WaitGroup, context.Context) {
@@ -76,6 +79,7 @@ func interruptHandler() (*sync.WaitGroup, context.Context) {
 	return &wg, ctx
 }
 
+//nolint:cyclop // function is already fairly clean
 func getHostInstance(ctx context.Context) (*crusoeapi.InstanceV1Alpha5, error) {
 	crusoeClient := crusoe.NewCrusoeClient(
 		viper.GetString(CrusoeAPIEndpointFlag),
@@ -84,28 +88,32 @@ func getHostInstance(ctx context.Context) (*crusoeapi.InstanceV1Alpha5, error) {
 		"crusoe-csi-driver/0.0.1",
 	)
 
-	fullNodeName := viper.GetString(NodeNameFlag)
-	nodeNameSplit := strings.SplitN(fullNodeName, ".", numExpectedComponents)
-	if len(nodeNameSplit) < 1 {
-		return nil, errInvalidNodeName
+	vmIDStringByteArray, err := os.ReadFile(vmIDFilePath)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %w", errVMIDReadFailed, err)
+	}
+
+	vmIDString := strings.TrimSpace(string(vmIDStringByteArray))
+	_, err = uuid.Parse(vmIDString)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %w", errVMIDParseFailed, err)
 	}
-	shortNodeName := nodeNameSplit[0]
 
 	var projectID string
 
 	projectID = viper.GetString(CrusoeProjectIDFlag)
 	if projectID == "" {
 		var ok bool
-		kubeClientConfig, err := rest.InClusterConfig()
-		if err != nil {
+		kubeClientConfig, configErr := rest.InClusterConfig()
+		if configErr != nil {
 			return nil, fmt.Errorf("could not get kube client config: %w", err)
 		}
 
-		kubeClient, err := kubernetes.NewForConfig(kubeClientConfig)
-		if err != nil {
+		kubeClient, clientErr := kubernetes.NewForConfig(kubeClientConfig)
+		if clientErr != nil {
 			return nil, fmt.Errorf("could not get kube client: %w", err)
 		}
-		hostNode, nodeFetchErr := kubeClient.CoreV1().Nodes().Get(ctx, fullNodeName, metav1.GetOptions{})
+		hostNode, nodeFetchErr := kubeClient.CoreV1().Nodes().Get(ctx, viper.GetString(NodeNameFlag), metav1.GetOptions{})
 		if nodeFetchErr != nil {
 			return nil, fmt.Errorf("could not fetch current node with kube client: %w", err)
 		}
@@ -118,16 +126,16 @@ func getHostInstance(ctx context.Context) (*crusoeapi.InstanceV1Alpha5, error) {
 
 	instances, _, err := crusoeClient.VMsApi.ListInstances(ctx, projectID,
 		&crusoeapi.VMsApiListInstancesOpts{
-			Names: optional.NewString(shortNodeName),
+			Ids: optional.NewString(vmIDString),
 		})
 	if err != nil {
 		return nil, fmt.Errorf("failed to list instances: %w", err)
 	}
 
 	if len(instances.Items) == 0 {
-		return nil, fmt.Errorf("%w: %s", errInstanceNotFound, shortNodeName)
+		return nil, fmt.Errorf("%w: %s", errInstanceNotFound, vmIDString)
 	} else if len(instances.Items) > 1 {
-		return nil, fmt.Errorf("%w: %s", errMultipleInstances, shortNodeName)
+		return nil, fmt.Errorf("%w: %s", errMultipleInstances, vmIDString)
 	}
 
 	return &instances.Items[0], nil