diff --git a/weed/command/mount_std.go b/weed/command/mount_std.go
index 44e945f23..3947a871a 100644
--- a/weed/command/mount_std.go
+++ b/weed/command/mount_std.go
@@ -91,7 +91,7 @@ func RunMount(option *MountOptions, umask os.FileMode) bool {
 
 	// detect mount folder mode
 	if *option.dirAutoCreate {
-		os.MkdirAll(dir, 0755)
+		os.MkdirAll(dir, os.FileMode(0777) &^ umask)
 	}
 	mountMode := os.ModeDir | 0755
 	fileInfo, err := os.Stat(dir)
diff --git a/weed/filer/meta_aggregator.go b/weed/filer/meta_aggregator.go
index 4918899ff..b90457339 100644
--- a/weed/filer/meta_aggregator.go
+++ b/weed/filer/meta_aggregator.go
@@ -78,7 +78,7 @@ func (ma *MetaAggregator) subscribeToOneFiler(f *Filer, self string, peer string
 		var counter int64
 		var synced bool
 		maybeReplicateMetadataChange = func(event *filer_pb.SubscribeMetadataResponse) {
-			if err := Replay(f.Store.ActualStore, event); err != nil {
+			if err := Replay(f.Store, event); err != nil {
 				glog.Errorf("failed to reply metadata change from %v: %v", peer, err)
 				return
 			}
diff --git a/weed/filesys/meta_cache/meta_cache.go b/weed/filesys/meta_cache/meta_cache.go
index ac193a493..06f634c72 100644
--- a/weed/filesys/meta_cache/meta_cache.go
+++ b/weed/filesys/meta_cache/meta_cache.go
@@ -17,7 +17,7 @@ import (
 // e.g. fill fileId field for chunks
 
 type MetaCache struct {
-	actualStore filer.FilerStore
+	localStore filer.FilerStore
 	sync.RWMutex
 	visitedBoundary *bounded_tree.BoundedTree
 	uidGidMapper    *UidGidMapper
@@ -25,7 +25,7 @@ type MetaCache struct {
 
 func NewMetaCache(dbFolder string, uidGidMapper *UidGidMapper) *MetaCache {
 	return &MetaCache{
-		actualStore:     openMetaStore(dbFolder),
+		localStore:      openMetaStore(dbFolder),
 		visitedBoundary: bounded_tree.NewBoundedTree(),
 		uidGidMapper:    uidGidMapper,
 	}
@@ -57,7 +57,7 @@ func (mc *MetaCache) InsertEntry(ctx context.Context, entry *filer.Entry) error
 
 func (mc *MetaCache) doInsertEntry(ctx context.Context, entry *filer.Entry) error {
 	filer_pb.BeforeEntrySerialization(entry.Chunks)
-	return mc.actualStore.InsertEntry(ctx, entry)
+	return mc.localStore.InsertEntry(ctx, entry)
 }
 
 func (mc *MetaCache) AtomicUpdateEntryFromFiler(ctx context.Context, oldPath util.FullPath, newEntry *filer.Entry) error {
@@ -71,7 +71,7 @@ func (mc *MetaCache) AtomicUpdateEntryFromFiler(ctx context.Context, oldPath uti
 				// skip the unnecessary deletion
 				// leave the update to the following InsertEntry operation
 			} else {
-				if err := mc.actualStore.DeleteEntry(ctx, oldPath); err != nil {
+				if err := mc.localStore.DeleteEntry(ctx, oldPath); err != nil {
 					return err
 				}
 			}
@@ -83,7 +83,7 @@ func (mc *MetaCache) AtomicUpdateEntryFromFiler(ctx context.Context, oldPath uti
 	if newEntry != nil {
 		newDir, _ := newEntry.DirAndName()
 		if mc.visitedBoundary.HasVisited(util.FullPath(newDir)) {
-			if err := mc.actualStore.InsertEntry(ctx, newEntry); err != nil {
+			if err := mc.localStore.InsertEntry(ctx, newEntry); err != nil {
 				return err
 			}
 		}
@@ -95,13 +95,13 @@ func (mc *MetaCache) UpdateEntry(ctx context.Context, entry *filer.Entry) error
 	mc.Lock()
 	defer mc.Unlock()
 	filer_pb.BeforeEntrySerialization(entry.Chunks)
-	return mc.actualStore.UpdateEntry(ctx, entry)
+	return mc.localStore.UpdateEntry(ctx, entry)
 }
 
 func (mc *MetaCache) FindEntry(ctx context.Context, fp util.FullPath) (entry *filer.Entry, err error) {
 	mc.RLock()
 	defer mc.RUnlock()
-	entry, err = mc.actualStore.FindEntry(ctx, fp)
+	entry, err = mc.localStore.FindEntry(ctx, fp)
 	if err != nil {
 		return nil, err
 	}
@@ -113,14 +113,14 @@ func (mc *MetaCache) FindEntry(ctx context.Context, fp util.FullPath) (entry *fi
 func (mc *MetaCache) DeleteEntry(ctx context.Context, fp util.FullPath) (err error) {
 	mc.Lock()
 	defer mc.Unlock()
-	return mc.actualStore.DeleteEntry(ctx, fp)
+	return mc.localStore.DeleteEntry(ctx, fp)
 }
 
 func (mc *MetaCache) ListDirectoryEntries(ctx context.Context, dirPath util.FullPath, startFileName string, includeStartFile bool, limit int) ([]*filer.Entry, error) {
 	mc.RLock()
 	defer mc.RUnlock()
 
-	entries, err := mc.actualStore.ListDirectoryEntries(ctx, dirPath, startFileName, includeStartFile, limit)
+	entries, err := mc.localStore.ListDirectoryEntries(ctx, dirPath, startFileName, includeStartFile, limit)
 	if err != nil {
 		return nil, err
 	}
@@ -134,7 +134,7 @@ func (mc *MetaCache) ListDirectoryEntries(ctx context.Context, dirPath util.Full
 func (mc *MetaCache) Shutdown() {
 	mc.Lock()
 	defer mc.Unlock()
-	mc.actualStore.Shutdown()
+	mc.localStore.Shutdown()
 }
 
 func (mc *MetaCache) mapIdFromFilerToLocal(entry *filer.Entry) {
diff --git a/weed/filesys/wfs.go b/weed/filesys/wfs.go
index 8d46e0862..0335d5d98 100644
--- a/weed/filesys/wfs.go
+++ b/weed/filesys/wfs.go
@@ -88,7 +88,7 @@ func NewSeaweedFileSystem(option *Option) *WFS {
 	cacheUniqueId := util.Md5String([]byte(option.FilerGrpcAddress + option.FilerMountRootPath + util.Version()))[0:4]
 	cacheDir := path.Join(option.CacheDir, cacheUniqueId)
 	if option.CacheSizeMB > 0 {
-		os.MkdirAll(cacheDir, 0755)
+		os.MkdirAll(cacheDir, os.FileMode(0777) &^ option.Umask)
 		wfs.chunkCache = chunk_cache.NewTieredChunkCache(256, cacheDir, option.CacheSizeMB)
 	}