[release-1.35] Fix restart of control-plane-only nodes attempting to reconcile from local datastore

pkg/cluster/bootstrap.go

@@ -287,7 +287,7 @@ func (c *Cluster) ReconcileBootstrapData(ctx context.Context, buf io.ReadSeeker,
 
 		storageClient, err := store.NewTemporaryStore(filepath.Join(c.config.DataDir, "db", "etcd"))
 		if err != nil {
-			return err
+			return pkgerrors.WithMessage(err, "failed to create temporary datastore client")
 		}
 		defer storageClient.Close()
 
@@ -465,7 +465,7 @@ func (c *Cluster) bootstrap(ctx context.Context) error {
 			logrus.Debugf("Failed to get bootstrap data from etcd proxy: %v", err)
 		} else {
 			if err := c.ReconcileBootstrapData(ctx, bytes.NewReader(data), &c.config.Runtime.ControlRuntimeBootstrap, false); err != nil {
-				logrus.Debugf("Failed to reconcile bootstrap data from etcd proxy: %v", err)
+				logrus.Debugf("Failed to reconcile with local datastore: %v", err)
 			} else {
 				return nil
 			}

pkg/etcd/etcd.go

@@ -718,6 +718,11 @@ func (e *ETCD) Register(handler http.Handler) (http.Handler, error) {
 // or if force is set to true, a new name will be generated and written to disk. The persistent
 // name is used on subsequent calls.
 func (e *ETCD) setName(force bool) error {
+	// don't create the name file if etcd is disabled
+	if e.config.DisableETCD {
+		return nil
+	}
+
 	fileName := nameFile(e.config)
 	data, err := os.ReadFile(fileName)
 	if os.IsNotExist(err) || force {
@@ -1085,6 +1090,11 @@ func addPort(address string, offset int) (string, error) {
 
 // RemovePeer removes a peer from the cluster. The peer name and IP address must both match.
 func (e *ETCD) RemovePeer(ctx context.Context, name, address string, allowSelfRemoval bool) error {
+	// do not remove self if we have never started etcd on this node
+	if name == "" {
+		return nil
+	}
+
 	ctx, cancel := context.WithTimeout(ctx, memberRemovalTimeout)
 	defer cancel()
 	members, err := e.client.MemberList(ctx)
@@ -1676,9 +1686,10 @@ func (e *ETCD) RemoveSelf(ctx context.Context) error {
 
 	// backup the data dir to avoid issues when re-enabling etcd
 	oldDataDir := dbDir(e.config) + "-old-" + strconv.Itoa(int(time.Now().Unix()))
-
-	// move the data directory to a temp path
-	return os.Rename(dbDir(e.config), oldDataDir)
+	if err := os.Rename(dbDir(e.config), oldDataDir); err != nil && !os.IsNotExist(err) {
+		return err
+	}
+	return nil
 }
 
 // DefaultEndpointConfig returns default kine endpoint config, with k3s default

pkg/etcd/store/store.go

@@ -9,6 +9,7 @@ import (
 
 	"github.com/k3s-io/kine/pkg/endpoint"
 	"github.com/otiai10/copy"
+	pkgerrors "github.com/pkg/errors"
 	"github.com/rancher/wrangler/v3/pkg/merr"
 	"github.com/sirupsen/logrus"
 	"go.etcd.io/etcd/api/v3/mvccpb"
@@ -61,7 +62,6 @@ func NewRemoteStore(config endpoint.ETCDConfig) (*RemoteStore, error) {
 	if err != nil {
 		return nil, err
 	}
-
 	c, err := clientv3.New(clientv3.Config{
 		Endpoints:   config.Endpoints,
 		DialTimeout: 5 * time.Second,
@@ -205,11 +205,19 @@ func NewStore(dataDir string) (*Store, error) {
 	}
 
 	cfg := config.ServerConfig{Logger: logger, DataDir: dataDir}
-	logrus.Infof("Opening etcd MVCC KV store at %s", cfg.BackendPath())
+	path := cfg.BackendPath()
+
+	// need to check for backend path ourselves, as backend.New just logs a panic
+	// via zap if it doesn't exist, which isn't fatal.
+	if _, err := os.Stat(path); err != nil {
+		return nil, pkgerrors.WithMessage(err, "failed to stat MVCC KV store backend path")
+	}
+
+	logrus.Infof("Opening etcd MVCC KV store at %s", path)
 
 	// open backend database
 	bcfg := backend.DefaultBackendConfig(logger)
-	bcfg.Path = cfg.BackendPath()
+	bcfg.Path = path
 	bcfg.UnsafeNoFsync = true
 	bcfg.BatchInterval = 0
 	bcfg.BatchLimit = 0