fix: add more optimizations to putMetacacheObject() (#12916)

- avoid extra lookup for 'xl.meta' since we are
  definitely sure that it doesn't exist.

- use this in newMultipartUpload() as well

- also additionally do not write with O_DSYNC
  to avoid loading the drives, instead create
  'xl.meta' for listing operations without
  O_DSYNC since these are ephemeral objects.

- do the same with newMultipartUpload() since
  it gets synced when the PutObjectPart() is
  attempted, we do not need to tax newMultipartUpload()
  instead.

cmd/erasure-object.go

@@ -677,11 +677,7 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 			onlineDisks[i] = nil
 			continue
 		}
-		if len(inlineBuffers) > 0 && inlineBuffers[i] != nil {
-			partsMetadata[i].Data = inlineBuffers[i].Bytes()
-		} else {
-			partsMetadata[i].Data = nil
-		}
+		partsMetadata[i].Data = inlineBuffers[i].Bytes()
 		partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize())
 		partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
 			PartNumber: 1,
@@ -695,9 +691,10 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 	// Fill all the necessary metadata.
 	// Update `xl.meta` content on each disks.
 	for index := range partsMetadata {
-		partsMetadata[index].Metadata = opts.UserDefined
 		partsMetadata[index].Size = n
+		partsMetadata[index].Fresh = true
 		partsMetadata[index].ModTime = modTime
+		partsMetadata[index].Metadata = opts.UserDefined
 	}
 
 	// Set an additional header when data is inlined.
@@ -714,33 +711,11 @@ func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *
 		}
 	}
 
-	g := errgroup.WithNErrs(len(onlineDisks))
-
-	// Rename file on all underlying storage disks.
-	for index := range onlineDisks {
-		index := index
-		g.Go(func() error {
-			if onlineDisks[index] == nil {
-				return errDiskNotFound
-			}
-			// Pick one FileInfo for a disk at index.
-			fi := partsMetadata[index]
-			// Assign index when index is initialized
-			if fi.Erasure.Index == 0 {
-				fi.Erasure.Index = index + 1
-			}
-
-			if fi.IsValid() {
-				return onlineDisks[index].WriteMetadata(ctx, minioMetaBucket, key, fi)
-			}
-			return errFileCorrupt
-		}, index)
+	if _, err = writeUniqueFileInfo(ctx, onlineDisks, minioMetaBucket, key, partsMetadata, writeQuorum); err != nil {
+		return ObjectInfo{}, toObjectErr(err, minioMetaBucket, key)
 	}
 
-	// Wait for all renames to finish.
-	errs := g.Wait()
-
-	return fi.ToObjectInfo(minioMetaBucket, key), reduceWriteQuorumErrs(ctx, errs, objectOpIgnoredErrs, writeQuorum)
+	return fi.ToObjectInfo(minioMetaBucket, key), nil
 }
 
 // PutObject - creates an object upon reading from the input stream