diff --git a/core/blockchain.go b/core/blockchain.go
index 90ee3a5fe3..6a07d91789 100644
--- a/core/blockchain.go
+++ b/core/blockchain.go
@@ -656,10 +656,10 @@ func (bc *BlockChain) setHeadBeyondRoot(head uint64, time uint64, root common.Ha
 	pivot := rawdb.ReadLastPivotNumber(bc.db)
 	frozen, _ := bc.db.Ancients()
 
-	// Rewind the blockchain, ensuring we don't end up with a stateless head
-	// block. Note, depth equality is permitted to allow using SetHead as a
-	// chain reparation mechanism without deleting any data!
 	updateFn := func(db ethdb.KeyValueWriter, header *types.Header) (*types.Header, bool) {
+		// Rewind the blockchain, ensuring we don't end up with a stateless head
+		// block. Note, depth equality is permitted to allow using SetHead as a
+		// chain reparation mechanism without deleting any data!
 		if currentBlock := bc.CurrentBlock(); currentBlock != nil && header.Number.Uint64() <= currentBlock.Number.Uint64() {
 			newHeadBlock := bc.GetBlock(header.Hash(), header.Number.Uint64())
 			if newHeadBlock == nil {
@@ -708,20 +708,21 @@ func (bc *BlockChain) setHeadBeyondRoot(head uint64, time uint64, root common.Ha
 			}
 			rawdb.WriteHeadBlockHash(db, newHeadBlock.Hash())
 
-			// The genesis state is missing, which is only possible in the path-based
-			// scheme. This situation occurs when the chain head is rewound below the
-			// pivot point. In this scenario, there is no possible recovery approach
-			// except for rerunning a snap sync. Do nothing here until the state syncer
-			// picks it up.
-			if newHeadBlock.NumberU64() == 0 && !bc.HasState(newHeadBlock.Root()) {
-				log.Info("Genesis state is missing, wait state sync")
-			}
 			// Degrade the chain markers if they are explicitly reverted.
 			// In theory we should update all in-memory markers in the
 			// last step, however the direction of SetHead is from high
 			// to low, so it's safe to update in-memory markers directly.
 			bc.currentBlock.Store(newHeadBlock.Header())
 			headBlockGauge.Update(int64(newHeadBlock.NumberU64()))
+
+			// The head state is missing, which is only possible in the path-based
+			// scheme. This situation occurs when the chain head is rewound below
+			// the pivot point. In this scenario, there is no possible recovery
+			// approach except for rerunning a snap sync. Do nothing here until the
+			// state syncer picks it up.
+			if !bc.HasState(newHeadBlock.Root()) {
+				log.Info("Chain is stateless, wait state sync", "number", newHeadBlock.Number(), "hash", newHeadBlock.Hash())
+			}
 		}
 		// Rewind the snap block in a simpleton way to the target head
 		if currentSnapBlock := bc.CurrentSnapBlock(); currentSnapBlock != nil && header.Number.Uint64() < currentSnapBlock.Number.Uint64() {
diff --git a/core/rawdb/accessors_sync.go b/core/rawdb/accessors_sync.go
index 0366f930c5..2dc08b3b72 100644
--- a/core/rawdb/accessors_sync.go
+++ b/core/rawdb/accessors_sync.go
@@ -78,22 +78,23 @@ func DeleteSkeletonHeader(db ethdb.KeyValueWriter, number uint64) {
 }
 
 const (
-	StateSyncing = uint8(1) // flags the state snap sync is not completed yet
-	StateSynced  = uint8(2) // flags the state snap sync is completed
+	StateSyncUnknown  = uint8(0) // flags the state snap sync is unknown
+	StateSyncRunning  = uint8(1) // flags the state snap sync is not completed yet
+	StateSyncFinished = uint8(2) // flags the state snap sync is completed
 )
 
 // ReadSnapSyncStatusFlag retrieves the state snap sync status flag.
 func ReadSnapSyncStatusFlag(db ethdb.KeyValueReader) uint8 {
-	blob, err := db.Get(syncStatusFlagKey)
+	blob, err := db.Get(snapSyncStatusFlagKey)
 	if err != nil || len(blob) != 1 {
-		return 0
+		return StateSyncUnknown
 	}
 	return blob[0]
 }
 
 // WriteSnapSyncStatusFlag stores the state snap sync status flag into database.
 func WriteSnapSyncStatusFlag(db ethdb.KeyValueWriter, flag uint8) {
-	if err := db.Put(syncStatusFlagKey, []byte{flag}); err != nil {
+	if err := db.Put(snapSyncStatusFlagKey, []byte{flag}); err != nil {
 		log.Crit("Failed to store sync status flag", "err", err)
 	}
 }
diff --git a/core/rawdb/database.go b/core/rawdb/database.go
index 123021cd6b..e97eeb2aa3 100644
--- a/core/rawdb/database.go
+++ b/core/rawdb/database.go
@@ -555,7 +555,7 @@ func InspectDatabase(db ethdb.Database, keyPrefix, keyStart []byte) error {
 				lastPivotKey, fastTrieProgressKey, snapshotDisabledKey, SnapshotRootKey, snapshotJournalKey,
 				snapshotGeneratorKey, snapshotRecoveryKey, txIndexTailKey, fastTxLookupLimitKey,
 				uncleanShutdownKey, badBlockKey, transitionStatusKey, skeletonSyncStatusKey,
-				persistentStateIDKey, trieJournalKey, snapshotSyncStatusKey, syncStatusFlagKey,
+				persistentStateIDKey, trieJournalKey, snapshotSyncStatusKey, snapSyncStatusFlagKey,
 			} {
 				if bytes.Equal(key, meta) {
 					metadata.Add(size)
diff --git a/core/rawdb/schema.go b/core/rawdb/schema.go
index 02d47e7629..91e4fbb974 100644
--- a/core/rawdb/schema.go
+++ b/core/rawdb/schema.go
@@ -91,8 +91,8 @@ var (
 	// transitionStatusKey tracks the eth2 transition status.
 	transitionStatusKey = []byte("eth2-transition")
 
-	// syncStatusFlagKey flags that status of state sync.
-	syncStatusFlagKey = []byte("sync-status")
+	// snapSyncStatusFlagKey flags that status of snap sync.
+	snapSyncStatusFlagKey = []byte("snapSync-status")
 
 	// Data item prefixes (use single byte to avoid mixing data types, avoid `i`, used for indexes).
 	headerPrefix       = []byte("h") // headerPrefix + num (uint64 big endian) + hash -> header
diff --git a/core/txpool/txpool.go b/core/txpool/txpool.go
index ae94bfa70e..1aed986c88 100644
--- a/core/txpool/txpool.go
+++ b/core/txpool/txpool.go
@@ -93,7 +93,9 @@ func New(gasTip *big.Int, chain BlockChain, subpools []SubPool) (*TxPool, error)
 		quit:         make(chan chan error),
 	}
 	if chain.HasState(head.Root) {
-		pool.init(gasTip, head)
+		if err := pool.init(gasTip, head); err != nil {
+			return nil, err
+		}
 		go pool.loop(head, chain)
 	} else {
 		go pool.lazyInit(gasTip, chain)
@@ -102,17 +104,18 @@ func New(gasTip *big.Int, chain BlockChain, subpools []SubPool) (*TxPool, error)
 }
 
 // init performs the initialization for subpools.
-func (p *TxPool) init(gasTip *big.Int, head *types.Header) {
+func (p *TxPool) init(gasTip *big.Int, head *types.Header) error {
 	for i, subpool := range p.subpools {
 		if err := subpool.Init(gasTip, head, p.reserver(i, subpool)); err != nil {
 			for j := i - 1; j >= 0; j-- {
 				p.subpools[j].Close()
 			}
-			// TODO(rjl493456442) can we shutdown the node gracefully?
-			log.Crit("Failed to initialize subpool", "err", err)
+			return err
 		}
 	}
 	p.inited.Store(true)
+	log.Info("Initialized subpools", "head", head.Number, "hash", head.Hash())
+	return nil
 }
 
 // lazyInit waits the signal that state sync is completed and initializes the subpools.
@@ -131,7 +134,10 @@ func (p *TxPool) lazyInit(gasTip *big.Int, chain BlockChain) {
 			if !chain.HasState(head.Root) {
 				continue // shouldn't happen
 			}
-			p.init(gasTip, head)
+			if err := p.init(gasTip, head); err != nil {
+				// TODO(rjl493456442) can we shutdown the node gracefully?
+				log.Crit("Failed to lazy init subpools", "err", err)
+			}
 			go p.loop(head, chain)
 			return
 
@@ -279,7 +285,7 @@ func (p *TxPool) loop(head *types.Header, chain BlockChain) {
 // new transaction, and drops all transactions below this threshold.
 func (p *TxPool) SetGasTip(tip *big.Int) {
 	if !p.inited.Load() {
-		log.Info("Skip tip adjustment as txpool hasn't been initialized")
+		log.Info("Skip tip adjustment as txpool hasn't been initialized", "provided", tip)
 		return
 	}
 	for _, subpool := range p.subpools {
diff --git a/eth/handler.go b/eth/handler.go
index 6254b3a01a..7f134d7677 100644
--- a/eth/handler.go
+++ b/eth/handler.go
@@ -172,10 +172,11 @@ func newHandler(config *handlerConfig) (*handler, error) {
 		head := h.chain.CurrentBlock()
 		if head.Number.Uint64() > 0 && h.chain.HasState(head.Root) {
 			// Print warning log if database is not empty to run snap sync.
-			log.Warn("Switch sync mode from snap sync to full sync")
+			log.Warn("Switch sync mode from snap sync to full sync", "reason", "snap sync complete")
 		} else {
 			// If snap sync was requested and our database is empty, grant it
 			h.snapSync.Store(true)
+			log.Info("Enabled snap sync", "head", head.Number, "hash", head.Hash())
 		}
 	}
 	// If the sync succeeds, mark the local node as synced and enable all features
diff --git a/eth/sync.go b/eth/sync.go
index 03357254e7..c7ba7c93d6 100644
--- a/eth/sync.go
+++ b/eth/sync.go
@@ -212,6 +212,7 @@ func (cs *chainSyncer) modeAndLocalHead() (downloader.SyncMode, *big.Int) {
 	if !cs.handler.chain.HasState(head.Root) {
 		block := cs.handler.chain.CurrentSnapBlock()
 		td := cs.handler.chain.GetTd(block.Hash(), block.Number.Uint64())
+		log.Info("Reenabled snap sync as chain is stateless")
 		return downloader.SnapSync, td
 	}
 	// Nope, we're really full syncing
diff --git a/trie/triedb/pathdb/database.go b/trie/triedb/pathdb/database.go
index 99259538f2..ad8db9045e 100644
--- a/trie/triedb/pathdb/database.go
+++ b/trie/triedb/pathdb/database.go
@@ -181,7 +181,7 @@ func New(diskdb ethdb.Database, config *Config) *Database {
 		}
 	}
 	// Disable database in case node is still in the initial state sync stage.
-	if rawdb.ReadSnapSyncStatusFlag(diskdb) == rawdb.StateSyncing && !db.readOnly {
+	if rawdb.ReadSnapSyncStatusFlag(diskdb) == rawdb.StateSyncRunning && !db.readOnly {
 		db.Deactivate()
 	}
 	log.Warn("Path-based state scheme is an experimental feature")
@@ -260,8 +260,8 @@ func (db *Database) Deactivate() error {
 	db.tree.bottom().markStale()
 
 	// Write the initial sync flag to persist it across restarts.
-	rawdb.WriteSnapSyncStatusFlag(db.diskdb, rawdb.StateSyncing)
-	log.Info("Disabled trie database")
+	rawdb.WriteSnapSyncStatusFlag(db.diskdb, rawdb.StateSyncRunning)
+	log.Info("Disabled trie database due to ongoing sync")
 	return nil
 }
 
@@ -304,7 +304,7 @@ func (db *Database) Activate(root common.Hash) error {
 
 	// Re-enable the database as the final step.
 	db.disabled = false
-	rawdb.WriteSnapSyncStatusFlag(db.diskdb, rawdb.StateSynced)
+	rawdb.WriteSnapSyncStatusFlag(db.diskdb, rawdb.StateSyncFinished)
 	log.Info("Rebuilt trie database", "root", root)
 	return nil
 }