fix: malfunctioning read repair and balancer

Author: buraksezer

internal/cluster/balancer/balancer.go

@@ -63,7 +63,7 @@ func (b *Balancer) isAlive() bool {
 	return true
 }
 
-func (b *Balancer) scanPartition(sign uint64, part *partitions.Partition, owners... discovery.Member) bool {
+func (b *Balancer) scanPartition(sign uint64, part *partitions.Partition, owners ...discovery.Member) bool {
 	var clean = true
 	part.Map().Range(func(name, tmp interface{}) bool {
 		u := tmp.(partitions.Fragment)
@@ -145,18 +145,18 @@ LOOP:
 		}
 
 		part := b.backup.PartitionByID(partID)
-		if part.Length() == 0 || part.OwnerCount() == 0{
+		if part.Length() == 0 || part.OwnerCount() == 0 {
 			continue
 		}
 
 		var (
-			counter = 1
+			counter       = 1
 			currentOwners []discovery.Member
 		)
 
 		owners := part.Owners()
 		for i := len(owners) - 1; i >= 0; i-- {
-			if counter >= b.config.ReplicaCount-1 {
+			if counter > b.config.ReplicaCount-1 {
 				break
 			}
 
@@ -173,6 +173,10 @@ LOOP:
 			currentOwners = append(currentOwners, owner)
 		}
 
+		if len(currentOwners) == 0 {
+			continue LOOP
+		}
+
 		if b.scanPartition(sign, part, currentOwners...) {
 			part.Map().Range(func(name, tmp interface{}) bool {
 				// Delete the moved storage unit instance. GC will free the allocated memory.

internal/cluster/routingtable/distribute.go

@@ -16,7 +16,6 @@ package routingtable
 
 import (
 	"errors"
-
 	"github.com/buraksezer/consistent"
 	"github.com/buraksezer/olric/internal/discovery"
 	"github.com/buraksezer/olric/internal/protocol"

internal/dmap/get.go

@@ -195,14 +195,15 @@ func (dm *DMap) sanitizeAndSortVersions(versions []*version) []*version {
 }
 
 func (dm *DMap) lookupOnReplicas(hkey uint64, key string) []*version {
-	var versions []*version
 	// Check backup.
 	backups := dm.s.backup.PartitionOwnersByHKey(hkey)
+	versions := make([]*version, 0, len(backups))
 	for _, replica := range backups {
 		req := protocol.NewDMapMessage(protocol.OpGetBackup)
 		req.SetDMap(dm.name)
 		req.SetKey(key)
-		ver := &version{host: &replica}
+		host := replica
+		v := &version{host: &host}
 		resp, err := dm.s.requestTo(replica.String(), req)
 		if err != nil {
 			if dm.s.log.V(3).Ok() {
@@ -212,21 +213,21 @@ func (dm *DMap) lookupOnReplicas(hkey uint64, key string) []*version {
 		} else {
 			data := dm.engine.NewEntry()
 			data.Decode(resp.Value())
-			ver.entry = data
+			v.entry = data
 		}
-		versions = append(versions, ver)
+		versions = append(versions, v)
 	}
 	return versions
 }
 
 func (dm *DMap) readRepair(winner *version, versions []*version) {
-	for _, ver := range versions {
-		if ver.entry != nil && winner.entry.Timestamp() == ver.entry.Timestamp() {
+	for _, version := range versions {
+		if version.entry != nil && winner.entry.Timestamp() == version.entry.Timestamp() {
 			continue
 		}
 
 		// Sync
-		tmp := *ver.host
+		tmp := *version.host
 		if tmp.CompareByID(dm.s.rt.This()) {
 			hkey := partitions.HKey(dm.name, winner.entry.Key())
 			f, err := dm.getOrCreateFragment(hkey, partitions.PRIMARY)
@@ -270,9 +271,9 @@ func (dm *DMap) readRepair(winner *version, versions []*version) {
 					TTL:       winner.entry.TTL(),
 				})
 			}
-			_, err := dm.s.requestTo(ver.host.String(), req)
+			_, err := dm.s.requestTo(version.host.String(), req)
 			if err != nil {
-				dm.s.log.V(3).Printf("[ERROR] Failed to synchronize replica %s: %v", ver.host, err)
+				dm.s.log.V(3).Printf("[ERROR] Failed to synchronize replica %s: %v", version.host, err)
 			}
 		}
 	}

internal/dmap/get_test.go

@@ -201,6 +201,7 @@ func TestDMap_Get_ReadRepair(t *testing.T) {
 	c2.ReplicaCount = 2
 	e2 := testcluster.NewEnvironment(c2)
 	s2 := cluster.AddMember(e2).(*Service)
+
 	defer cluster.Shutdown()
 
 	// Call DMap.Put on S1
@@ -230,7 +231,6 @@ func TestDMap_Get_ReadRepair(t *testing.T) {
 	c3.ReplicaCount = 2
 	e3 := testcluster.NewEnvironment(c3)
 	s3 := cluster.AddMember(e3).(*Service)
-	defer cluster.Shutdown()
 
 	// Call DMap.Get on S2
 	dm2, err := s3.NewDMap("mymap")