fix: scanner in healing cycle must use actual size (#16589)

cmd/erasure-healing.go

@@ -411,6 +411,11 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 			erasure.ShardFileSize(latestMeta.Parts[0].ActualSize) < smallFileThreshold)
 	}
 
+	result.ObjectSize, err = latestMeta.GetActualSize()
+	if err != nil {
+		return result, err
+	}
+
 	// Loop to find number of disks with valid data, per-drive
 	// data state and a list of outdated disks on which data needs
 	// to be healed.
@@ -421,9 +426,6 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 		switch {
 		case v != nil:
 			driveState = madmin.DriveStateOk
-			// If data is sane on any one disk, we can
-			// extract the correct object size.
-			result.ObjectSize = partsMetadata[i].Size
 		case errs[i] == errDiskNotFound, dataErrs[i] == errDiskNotFound:
 			driveState = madmin.DriveStateOffline
 		case errs[i] == errFileNotFound, errs[i] == errFileVersionNotFound, errs[i] == errVolumeNotFound:
@@ -673,9 +675,6 @@ func (er *erasureObjects) healObject(ctx context.Context, bucket string, object
 		})
 	}
 
-	// Set the size of the object in the heal result
-	result.ObjectSize = latestMeta.Size
-
 	return result, nil
 }
 

cmd/erasure-metadata.go

@@ -23,15 +23,18 @@ import (
 	"fmt"
 	"net/http"
 	"sort"
+	"strconv"
 	"strings"
 	"time"
 
 	"github.com/minio/minio/internal/amztime"
 	"github.com/minio/minio/internal/bucket/replication"
+	"github.com/minio/minio/internal/crypto"
 	"github.com/minio/minio/internal/hash/sha256"
 	xhttp "github.com/minio/minio/internal/http"
 	"github.com/minio/minio/internal/logger"
 	"github.com/minio/minio/internal/sync/errgroup"
+	"github.com/minio/sio"
 )
 
 // Object was stored with additional erasure codes due to degraded system at upload time
@@ -104,6 +107,52 @@ func (fi FileInfo) IsValid() bool {
 		correctIndexes)
 }
 
+func (fi FileInfo) checkMultipart() (int64, bool) {
+	if len(fi.Parts) == 0 {
+		return 0, false
+	}
+	if !crypto.IsMultiPart(fi.Metadata) {
+		return 0, false
+	}
+	var size int64
+	for _, part := range fi.Parts {
+		psize, err := sio.DecryptedSize(uint64(part.Size))
+		if err != nil {
+			return 0, false
+		}
+		size += int64(psize)
+	}
+
+	return size, len(extractETag(fi.Metadata)) != 32
+}
+
+// GetActualSize - returns the actual size of the stored object
+func (fi FileInfo) GetActualSize() (int64, error) {
+	if _, ok := fi.Metadata[ReservedMetadataPrefix+"compression"]; ok {
+		sizeStr, ok := fi.Metadata[ReservedMetadataPrefix+"actual-size"]
+		if !ok {
+			return -1, errInvalidDecompressedSize
+		}
+		size, err := strconv.ParseInt(sizeStr, 10, 64)
+		if err != nil {
+			return -1, errInvalidDecompressedSize
+		}
+		return size, nil
+	}
+	if _, ok := crypto.IsEncrypted(fi.Metadata); ok {
+		size, ok := fi.checkMultipart()
+		if !ok {
+			size, err := sio.DecryptedSize(uint64(fi.Size))
+			if err != nil {
+				err = errObjectTampered // assign correct error type
+			}
+			return int64(size), err
+		}
+		return size, nil
+	}
+	return fi.Size, nil
+}
+
 // ToObjectInfo - Converts metadata to object info.
 func (fi FileInfo) ToObjectInfo(bucket, object string, versioned bool) ObjectInfo {
 	object = decodeDirObject(object)