Enforce quota checks after crawl (#10036)

Enforce bucket quotas when crawling has finished. 
This ensures that we will not do quota enforcement on old data.

Additionally, delete less if we are closer to quota than we thought.

cmd/bucket-quota.go

@@ -22,9 +22,7 @@ import (
 	"fmt"
 	"time"
 
-	"github.com/minio/minio/cmd/config"
 	"github.com/minio/minio/cmd/logger"
-	"github.com/minio/minio/pkg/env"
 	"github.com/minio/minio/pkg/event"
 	"github.com/minio/minio/pkg/madmin"
 )
@@ -120,149 +118,112 @@ func enforceBucketQuota(ctx context.Context, bucket string, size int64) error {
 	return globalBucketQuotaSys.check(ctx, bucket, size)
 }
 
-const (
-	bgQuotaInterval = 1 * time.Hour
-)
-
-// initQuotaEnforcement starts the routine that deletes objects in bucket
-// that exceeds the FIFO quota
-func initQuotaEnforcement(ctx context.Context, objAPI ObjectLayer) {
-	if env.Get(envDataUsageCrawlConf, config.EnableOn) == config.EnableOn {
-		go startBucketQuotaEnforcement(ctx, objAPI)
+// enforceFIFOQuota deletes objects in FIFO order until sufficient objects
+// have been deleted so as to bring bucket usage within quota.
+func enforceFIFOQuotaBucket(ctx context.Context, objectAPI ObjectLayer, bucket string, bui BucketUsageInfo) {
+	// Check if the current bucket has quota restrictions, if not skip it
+	cfg, err := globalBucketQuotaSys.Get(bucket)
+	if err != nil {
+		return
 	}
-}
 
-func startBucketQuotaEnforcement(ctx context.Context, objAPI ObjectLayer) {
-	for {
-		select {
-		case <-ctx.Done():
-			return
-		case <-time.NewTimer(bgQuotaInterval).C:
-			enforceFIFOQuota(ctx, objAPI)
-		}
+	if cfg.Type != madmin.FIFOQuota {
+		return
+	}
 
+	var toFree uint64
+	if bui.Size > cfg.Quota && cfg.Quota > 0 {
+		toFree = bui.Size - cfg.Quota
 	}
-}
 
-// enforceFIFOQuota deletes objects in FIFO order until sufficient objects
-// have been deleted so as to bring bucket usage within quota
-func enforceFIFOQuota(ctx context.Context, objectAPI ObjectLayer) {
-	// Turn off quota enforcement if data usage info is unavailable.
-	if env.Get(envDataUsageCrawlConf, config.EnableOn) == config.EnableOff {
+	if toFree <= 0 {
 		return
 	}
 
-	buckets, err := objectAPI.ListBuckets(ctx)
-	if err != nil {
+	// Allocate new results channel to receive ObjectInfo.
+	objInfoCh := make(chan ObjectInfo)
+
+	versioned := globalBucketVersioningSys.Enabled(bucket)
+
+	// Walk through all objects
+	if err := objectAPI.Walk(ctx, bucket, "", objInfoCh, ObjectOptions{WalkVersions: versioned}); err != nil {
 		logger.LogIf(ctx, err)
 		return
 	}
 
-	dataUsageInfo, err := loadDataUsageFromBackend(ctx, objectAPI)
+	// reuse the fileScorer used by disk cache to score entries by
+	// ModTime to find the oldest objects in bucket to delete. In
+	// the context of bucket quota enforcement - number of hits are
+	// irrelevant.
+	scorer, err := newFileScorer(toFree, time.Now().Unix(), 1)
 	if err != nil {
 		logger.LogIf(ctx, err)
 		return
 	}
 
-	for _, binfo := range buckets {
-		bucket := binfo.Name
-
-		bui, ok := dataUsageInfo.BucketsUsage[bucket]
-		if !ok {
-			// bucket doesn't exist anymore, or we
-			// do not have any information to proceed.
-			continue
-		}
-
-		// Check if the current bucket has quota restrictions, if not skip it
-		cfg, err := globalBucketQuotaSys.Get(bucket)
-		if err != nil {
-			continue
-		}
-
-		if cfg.Type != madmin.FIFOQuota {
+	rcfg, _ := globalBucketObjectLockSys.Get(bucket)
+	for obj := range objInfoCh {
+		if obj.DeleteMarker {
+			// Delete markers are automatically added for FIFO purge.
+			scorer.addFileWithObjInfo(obj, 1)
 			continue
 		}
-
-		var toFree uint64
-		if bui.Size > cfg.Quota && cfg.Quota > 0 {
-			toFree = bui.Size - cfg.Quota
-		}
-
-		if toFree == 0 {
+		// skip objects currently under retention
+		if rcfg.LockEnabled && enforceRetentionForDeletion(ctx, obj) {
 			continue
 		}
+		scorer.addFileWithObjInfo(obj, 1)
+	}
 
-		// Allocate new results channel to receive ObjectInfo.
-		objInfoCh := make(chan ObjectInfo)
-
-		versioned := globalBucketVersioningSys.Enabled(bucket)
-
-		// Walk through all objects
-		if err := objectAPI.Walk(ctx, bucket, "", objInfoCh, ObjectOptions{WalkVersions: versioned}); err != nil {
-			logger.LogIf(ctx, err)
-			continue
+	// If we saw less than quota we are good.
+	if scorer.seenBytes <= cfg.Quota {
+		return
+	}
+	// Calculate how much we want to delete now.
+	toFreeNow := scorer.seenBytes - cfg.Quota
+	// We were less over quota than we thought. Adjust so we delete less.
+	// If we are more over, leave it for the next run to pick up.
+	if toFreeNow < toFree {
+		if !scorer.adjustSaveBytes(int64(toFreeNow) - int64(toFree)) {
+			// We got below or at quota.
+			return
 		}
+	}
 
-		// reuse the fileScorer used by disk cache to score entries by
-		// ModTime to find the oldest objects in bucket to delete. In
-		// the context of bucket quota enforcement - number of hits are
-		// irrelevant.
-		scorer, err := newFileScorer(toFree, time.Now().Unix(), 1)
-		if err != nil {
-			logger.LogIf(ctx, err)
+	var objects []ObjectToDelete
+	numKeys := len(scorer.fileObjInfos())
+	for i, obj := range scorer.fileObjInfos() {
+		objects = append(objects, ObjectToDelete{
+			ObjectName: obj.Name,
+			VersionID:  obj.VersionID,
+		})
+		if len(objects) < maxDeleteList && (i < numKeys-1) {
+			// skip deletion until maxDeleteList or end of slice
 			continue
 		}
 
-		rcfg, _ := globalBucketObjectLockSys.Get(bucket)
-		for obj := range objInfoCh {
-			if obj.DeleteMarker {
-				// Delete markers are automatically added for FIFO purge.
-				scorer.addFileWithObjInfo(obj, 1)
-				continue
-			}
-			// skip objects currently under retention
-			if rcfg.LockEnabled && enforceRetentionForDeletion(ctx, obj) {
-				continue
-			}
-			scorer.addFileWithObjInfo(obj, 1)
+		if len(objects) == 0 {
+			break
 		}
 
-		var objects []ObjectToDelete
-		numKeys := len(scorer.fileObjInfos())
-		for i, obj := range scorer.fileObjInfos() {
-			objects = append(objects, ObjectToDelete{
-				ObjectName: obj.Name,
-				VersionID:  obj.VersionID,
-			})
-			if len(objects) < maxDeleteList && (i < numKeys-1) {
-				// skip deletion until maxDeleteList or end of slice
+		// Deletes a list of objects.
+		_, deleteErrs := objectAPI.DeleteObjects(ctx, bucket, objects, ObjectOptions{
+			Versioned: versioned,
+		})
+		for i := range deleteErrs {
+			if deleteErrs[i] != nil {
+				logger.LogIf(ctx, deleteErrs[i])
 				continue
 			}
 
-			if len(objects) == 0 {
-				break
-			}
-
-			// Deletes a list of objects.
-			_, deleteErrs := objectAPI.DeleteObjects(ctx, bucket, objects, ObjectOptions{
-				Versioned: versioned,
+			// Notify object deleted event.
+			sendEvent(eventArgs{
+				EventName:  event.ObjectRemovedDelete,
+				BucketName: bucket,
+				Object:     obj,
+				Host:       "Internal: [FIFO-QUOTA-EXPIRY]",
 			})
-			for i := range deleteErrs {
-				if deleteErrs[i] != nil {
-					logger.LogIf(ctx, deleteErrs[i])
-					continue
-				}
-
-				// Notify object deleted event.
-				sendEvent(eventArgs{
-					EventName:  event.ObjectRemovedDelete,
-					BucketName: bucket,
-					Object:     obj,
-					Host:       "Internal: [FIFO-QUOTA-EXPIRY]",
-				})
-			}
-			objects = nil
 		}
+		objects = nil
 	}
 }

cmd/data-usage-cache.go

@@ -335,6 +335,21 @@ func (d *dataUsageCache) bucketsUsageInfo(buckets []BucketInfo) map[string]Bucke
 	return dst
 }
 
+// bucketUsageInfo returns the buckets usage info.
+// If not found all values returned are zero values.
+func (d *dataUsageCache) bucketUsageInfo(bucket string) BucketUsageInfo {
+	e := d.find(bucket)
+	if e == nil {
+		return BucketUsageInfo{}
+	}
+	flat := d.flatten(*e)
+	return BucketUsageInfo{
+		Size:                 uint64(flat.Size),
+		ObjectsCount:         uint64(flat.Objects),
+		ObjectSizesHistogram: flat.ObjSizes.toMap(),
+	}
+}
+
 // sizeRecursive returns the path as a flattened entry.
 func (d *dataUsageCache) sizeRecursive(path string) *dataUsageEntry {
 	root := d.find(path)

cmd/disk-cache-utils.go

@@ -295,6 +295,7 @@ type fileScorer struct {
 	// The list is kept sorted according to score, highest at top, lowest at bottom.
 	queue       list.List
 	queuedBytes uint64
+	seenBytes   uint64
 }
 
 type queuedFile struct {
@@ -337,6 +338,7 @@ func (f *fileScorer) addFileWithObjInfo(objInfo ObjectInfo, hits int) {
 		versionID: objInfo.VersionID,
 		size:      uint64(objInfo.Size),
 	}
+	f.seenBytes += uint64(objInfo.Size)
 
 	var score float64
 	if objInfo.ModTime.IsZero() {
@@ -369,9 +371,14 @@ func (f *fileScorer) addFileWithObjInfo(objInfo ObjectInfo, hits int) {
 
 // adjustSaveBytes allows to adjust the number of bytes to save.
 // This can be used to adjust the count on the fly.
-// Returns true if there still is a need to delete files (saveBytes >0),
+// Returns true if there still is a need to delete files (n+saveBytes >0),
 // false if no more bytes needs to be saved.
 func (f *fileScorer) adjustSaveBytes(n int64) bool {
+	if int64(f.saveBytes)+n <= 0 {
+		f.saveBytes = 0
+		f.trimQueue()
+		return false
+	}
 	if n < 0 {
 		f.saveBytes -= ^uint64(n - 1)
 	} else {

cmd/erasure-zones.go

@@ -337,13 +337,15 @@ func (z *erasureZones) CrawlAndGetDataUsage(ctx context.Context, bf *bloomFilter
 		updateTicker := time.NewTicker(30 * time.Second)
 		defer updateTicker.Stop()
 		var lastUpdate time.Time
+
+		// We need to merge since we will get the same buckets from each zone.
+		// Therefore to get the exact bucket sizes we must merge before we can convert.
+		allMerged := dataUsageCache{Info: dataUsageCacheInfo{Name: dataUsageRoot}}
+
 		update := func() {
 			mu.Lock()
 			defer mu.Unlock()
 
-			// We need to merge since we will get the same buckets from each zone.
-			// Therefore to get the exact bucket sizes we must merge before we can convert.
-			allMerged := dataUsageCache{Info: dataUsageCacheInfo{Name: dataUsageRoot}}
 			for _, info := range results {
 				if info.Info.LastUpdate.IsZero() {
 					// Not filled yet.
@@ -362,6 +364,10 @@ func (z *erasureZones) CrawlAndGetDataUsage(ctx context.Context, bf *bloomFilter
 				return
 			case v := <-updateCloser:
 				update()
+				// Enforce quotas when all is done.
+				for _, b := range allBuckets {
+					enforceFIFOQuotaBucket(ctx, z, b.Name, allMerged.bucketUsageInfo(b.Name))
+				}
 				close(v)
 				return
 			case <-updateTicker.C:

cmd/fs-v1.go

@@ -301,6 +301,7 @@ func (fs *FSObjects) CrawlAndGetDataUsage(ctx context.Context, bf *bloomFilter,
 		logger.LogIf(ctx, totalCache.save(ctx, fs, dataUsageCacheName))
 		cloned := totalCache.clone()
 		updates <- cloned.dui(dataUsageRoot, buckets)
+		enforceFIFOQuotaBucket(ctx, fs, b.Name, cloned.bucketUsageInfo(b.Name))
 	}
 
 	return nil

cmd/server-main.go

@@ -360,7 +360,6 @@ func startBackgroundOps(ctx context.Context, objAPI ObjectLayer) {
 	}
 
 	initDataCrawler(ctx, objAPI)
-	initQuotaEnforcement(ctx, objAPI)
 }
 
 // serverMain handler called for 'minio server' command.