do advance at start

store/store.go

@@ -376,14 +376,17 @@ func (s *Store[H]) Append(ctx context.Context, headers ...H) error {
 // (1) Appends not to be blocked on long disk IO writes and underlying DB compactions
 // (2) Batching header writes
 func (s *Store[H]) flushLoop() {
+	// advance based on what we have on disk.
+	s.doAdvanceContiguousHead(context.Background(), s.Height())
+
 	defer close(s.writesDn)
 	ctx := context.Background()
 	for headers := range s.writes {
 		// add headers to the pending and ensure they are accessible
 		s.pending.Append(headers...)
 		// try to advance contiguousHead if we don't have gaps.
 		// and notify waiters in heightSub.
-		s.advanceContiguousHead(ctx, headers...)
+		s.tryAdvanceContiguousHead(ctx, headers...)
 		// don't flush and continue if pending batch is not grown enough,
 		// and Store is not stopping(headers == nil)
 		if s.pending.Len() < s.Params.WriteBatchSize && headers != nil {
@@ -497,17 +500,19 @@ func (s *Store[H]) get(ctx context.Context, hash header.Hash) ([]byte, error) {
 }
 
 // try advance contiguous head based on already written headers.
-func (s *Store[H]) advanceContiguousHead(ctx context.Context, headers ...H) {
+func (s *Store[H]) tryAdvanceContiguousHead(ctx context.Context, headers ...H) {
 	// always inform heightSub about new headers seen
 	for _, h := range headers {
 		s.heightSub.UnblockHeight(h.Height())
 	}
 
 	currHead := s.contiguousHead.Load()
-	if currHead == nil {
-		return
+	if currHead != nil {
+		s.doAdvanceContiguousHead(ctx, (*currHead).Height())
 	}
-	currHeight := (*currHead).Height()
+}
+
+func (s *Store[H]) doAdvanceContiguousHead(ctx context.Context, currHeight uint64) {
 	prevHeight := currHeight
 
 	// TODO(cristaloleg): benchmark this timeout or make it dynamic.