diff --git a/CHANGELOG_PENDING.md b/CHANGELOG_PENDING.md
index ae38af93745..151b15d0a3b 100644
--- a/CHANGELOG_PENDING.md
+++ b/CHANGELOG_PENDING.md
@@ -27,3 +27,4 @@
   now been fixed
 - [p2p] \#9500 prevent peers who have errored being added to the peer_set (@jmalicevic)
 - [indexer] \#9473 fix bug that caused the psql indexer to index empty blocks whenever one of the transactions returned a non zero code. The relevant deduplication logic has been moved within the kv indexer only (@cmwaters)
+- [blocksync] \#9518 handle the case when the sending queue is full: retry block request after a timeout
diff --git a/blockchain/v0/pool.go b/blockchain/v0/pool.go
index 1328bfa2704..e3923b0d9bc 100644
--- a/blockchain/v0/pool.go
+++ b/blockchain/v0/pool.go
@@ -32,6 +32,7 @@ const (
 	maxTotalRequesters        = 600
 	maxPendingRequests        = maxTotalRequesters
 	maxPendingRequestsPerPeer = 20
+	requestRetrySeconds       = 30
 
 	// Minimum recv rate to ensure we're receiving blocks from a peer fast
 	// enough. If a peer is not sending us data at at least that rate, we
@@ -602,7 +603,7 @@ OUTER_LOOP:
 			}
 			peer = bpr.pool.pickIncrAvailablePeer(bpr.height)
 			if peer == nil {
-				// log.Info("No peers available", "height", height)
+				bpr.Logger.Debug("No peers currently available; will retry shortly", "height", bpr.height)
 				time.Sleep(requestIntervalMS * time.Millisecond)
 				continue PICK_PEER_LOOP
 			}
@@ -612,6 +613,7 @@ OUTER_LOOP:
 		bpr.peerID = peer.id
 		bpr.mtx.Unlock()
 
+		to := time.NewTimer(requestRetrySeconds * time.Second)
 		// Send request and wait.
 		bpr.pool.sendRequest(bpr.height, peer.id)
 	WAIT_LOOP:
@@ -624,6 +626,11 @@ OUTER_LOOP:
 				return
 			case <-bpr.Quit():
 				return
+			case <-to.C:
+				bpr.Logger.Debug("Retrying block request after timeout", "height", bpr.height, "peer", bpr.peerID)
+				// Simulate a redo
+				bpr.reset()
+				continue OUTER_LOOP
 			case peerID := <-bpr.redoCh:
 				if peerID == bpr.peerID {
 					bpr.reset()