cap MaxNonceGap to 4, add delay between batch messages during republish

chain/messagepool/messagepool.go

@@ -7,7 +7,6 @@ import (
 	"fmt"
 	"math"
 	stdbig "math/big"
-	"runtime"
 	"sort"
 	"sync"
 	"time"
@@ -53,7 +52,7 @@ var minimumBaseFee = types.NewInt(uint64(build.MinimumBaseFee))
 
 var MaxActorPendingMessages = 1000
 
-var MaxNonceGap = uint64(16)
+var MaxNonceGap = uint64(4)
 
 var (
 	ErrMessageTooBig = errors.New("message too big")
@@ -82,13 +81,6 @@ const (
 	localUpdates = "update"
 )
 
-func init() {
-	numcpus := uint64(runtime.NumCPU())
-	if numcpus < MaxNonceGap {
-		MaxNonceGap = numcpus
-	}
-}
-
 type MessagePool struct {
 	lk sync.Mutex
 

chain/messagepool/repub.go

@@ -3,6 +3,7 @@ package messagepool
 import (
 	"context"
 	"sort"
+	"time"
 
 	"golang.org/x/xerrors"
 
@@ -15,6 +16,8 @@ import (
 
 const repubMsgLimit = 30
 
+var RepublishBatchDelay = 200 * time.Millisecond
+
 func (mp *MessagePool) republishPendingMessages() error {
 	mp.curTsLk.Lock()
 	ts := mp.curTs
@@ -131,6 +134,12 @@ func (mp *MessagePool) republishPendingMessages() error {
 		}
 
 		count++
+
+		if count < len(msgs) {
+			// this delay is here to encourage the pubsub subsystem to process the messages serially
+			// and avoid creating nonce gaps because of concurrent validation.
+			time.Sleep(RepublishBatchDelay)
+		}
 	}
 
 	// track most recently republished messages

chain/messagepool/repub_test.go

@@ -12,6 +12,12 @@ import (
 )
 
 func TestRepubMessages(t *testing.T) {
+	oldRepublishBatchDelay = RepublishBatchDelay
+	RepublishBatchDelay = time.Microsecond
+	defer func() {
+		RepublishBatchDelay = oldRepublishBatchDelay
+	}()
+
 	tma := newTestMpoolAPI()
 	ds := datastore.NewMapDatastore()