Enhancements to daily-sweeper routine to reduce CPU load (#8209)

- ListObjectsHeal should list only objects
  which need healing, not the entire namespace.
- DeleteObjects() to be used to delete 1000s of
  objects in bulk instead of serially.

cmd/daily-lifecycle-ops.go

@@ -139,22 +139,24 @@ func lifecycleRound(ctx context.Context, objAPI ObjectLayer) error {
 			if err != nil {
 				continue
 			}
+			var objects []string
 			for _, obj := range res.Objects {
 				// Find the action that need to be executed
 				action := l.ComputeAction(obj.Name, obj.ModTime)
 				switch action {
 				case lifecycle.DeleteAction:
-					objAPI.DeleteObject(ctx, bucket.Name, obj.Name)
+					objects = append(objects, obj.Name)
 				default:
-					// Nothing
-
+					// Do nothing, for now.
 				}
 			}
+			// Deletes a list of objects.
+			objAPI.DeleteObjects(ctx, bucket.Name, objects)
 			if !res.IsTruncated {
+				// We are done here, proceed to next bucket.
 				break
-			} else {
-				marker = res.NextMarker
 			}
+			marker = res.NextMarker
 		}
 	}
 

cmd/daily-sweeper.go

@@ -91,6 +91,7 @@ func sweepRound(ctx context.Context, objAPI ObjectLayer) error {
 			if err != nil {
 				continue
 			}
+
 			for _, obj := range res.Objects {
 				for _, l := range copyDailySweepListeners() {
 					l <- pathJoin(bucket.Name, obj.Name)
@@ -133,7 +134,6 @@ func dailySweeper() {
 	// Start with random sleep time, so as to avoid "synchronous checks" between servers
 	time.Sleep(time.Duration(rand.Float64() * float64(time.Hour)))
 
-	// Perform a sweep round each month
 	for {
 		if time.Since(lastSweepTime) < 30*24*time.Hour {
 			time.Sleep(time.Hour)
@@ -147,13 +147,12 @@ func dailySweeper() {
 			// instance doing the sweep round
 			case OperationTimedOut:
 				lastSweepTime = time.Now()
-			default:
-				logger.LogIf(ctx, err)
-				time.Sleep(time.Minute)
 				continue
 			}
-		} else {
-			lastSweepTime = time.Now()
+			logger.LogIf(ctx, err)
+			time.Sleep(time.Minute)
+			continue
 		}
+		lastSweepTime = time.Now()
 	}
 }

cmd/xl-sets.go

@@ -823,7 +823,7 @@ func (f *FileInfoCh) Push(fi FileInfo) {
 
 // Calculate least entry across multiple FileInfo channels, additionally
 // returns a boolean to indicate if the caller needs to call again.
-func leastEntry(entriesCh []FileInfoCh, readQuorum int) (FileInfo, bool) {
+func leastEntry(entriesCh []FileInfoCh, totalDrives int, heal bool) (FileInfo, bool) {
 	var entriesValid = make([]bool, len(entriesCh))
 	var entries = make([]FileInfo, len(entriesCh))
 	for i := range entriesCh {
@@ -879,27 +879,30 @@ func leastEntry(entriesCh []FileInfoCh, readQuorum int) (FileInfo, bool) {
 		entriesCh[i].Push(entries[i])
 	}
 
-	if readQuorum < 0 {
-		return lentry, isTruncated
-	}
-
 	quorum := lentry.Quorum
 	if quorum == 0 {
-		quorum = readQuorum
+		quorum = totalDrives / 2
 	}
 
-	if leastEntryCount >= quorum {
-		return lentry, isTruncated
+	if heal {
+		// When healing is enabled, we should
+		// list only objects which need healing.
+		if leastEntryCount != totalDrives {
+			return lentry, isTruncated
+		}
+	} else {
+		if leastEntryCount >= quorum {
+			return lentry, isTruncated
+		}
 	}
-
-	return leastEntry(entriesCh, readQuorum)
+	return leastEntry(entriesCh, totalDrives, heal)
 }
 
 // mergeEntriesCh - merges FileInfo channel to entries upto maxKeys.
-func mergeEntriesCh(entriesCh []FileInfoCh, maxKeys int, readQuorum int) (entries FilesInfo) {
+func mergeEntriesCh(entriesCh []FileInfoCh, maxKeys int, totalDrives int, heal bool) (entries FilesInfo) {
 	var i = 0
 	for {
-		fi, valid := leastEntry(entriesCh, readQuorum)
+		fi, valid := leastEntry(entriesCh, totalDrives, heal)
 		if !valid {
 			break
 		}
@@ -948,7 +951,6 @@ func (s *xlSets) listObjectsNonSlash(ctx context.Context, bucket, prefix, marker
 	recursive := true
 	entryChs := s.startMergeWalks(context.Background(), bucket, prefix, "", recursive, endWalkCh)
 
-	readQuorum := s.drivesPerSet / 2
 	var objInfos []ObjectInfo
 	var eof bool
 	var prevPrefix string
@@ -957,7 +959,7 @@ func (s *xlSets) listObjectsNonSlash(ctx context.Context, bucket, prefix, marker
 		if len(objInfos) == maxKeys {
 			break
 		}
-		result, ok := leastEntry(entryChs, readQuorum)
+		result, ok := leastEntry(entryChs, s.drivesPerSet, false)
 		if !ok {
 			eof = true
 			break
@@ -1087,12 +1089,7 @@ func (s *xlSets) listObjects(ctx context.Context, bucket, prefix, marker, delimi
 		entryChs = s.startMergeWalks(context.Background(), bucket, prefix, marker, recursive, endWalkCh)
 	}
 
-	readQuorum := s.drivesPerSet / 2
-	if heal {
-		readQuorum = -1
-	}
-
-	entries := mergeEntriesCh(entryChs, maxKeys, readQuorum)
+	entries := mergeEntriesCh(entryChs, maxKeys, s.drivesPerSet, heal)
 	if len(entries.Files) == 0 {
 		return loi, nil
 	}