Make sure to close the disk connections (#5752)

Since we do not re-use storageDisks after moving
the connections to object layer we should close them
appropriately otherwise we have a lot of connection
leaks and these can compound as the time goes by.

This PR also refactors the initialization code to
re-use storageDisks for given set of endpoints until
we have confirmed a valid reference format.

cmd/format-xl.go

@@ -315,27 +315,18 @@ func shouldInitXLDisks(errs []error) bool {
 }
 
 // loadFormatXLAll - load all format config from all input disks in parallel.
-func loadFormatXLAll(endpoints EndpointList) ([]*formatXLV3, []error) {
+func loadFormatXLAll(storageDisks []StorageAPI) ([]*formatXLV3, []error) {
 	// Initialize sync waitgroup.
 	var wg = &sync.WaitGroup{}
 
-	bootstrapDisks := make([]StorageAPI, len(endpoints))
-	for i, endpoint := range endpoints {
-		disk, err := newStorageAPI(endpoint)
-		if err != nil {
-			continue
-		}
-		bootstrapDisks[i] = disk
-	}
-
 	// Initialize list of errors.
-	var sErrs = make([]error, len(bootstrapDisks))
+	var sErrs = make([]error, len(storageDisks))
 
 	// Initialize format configs.
-	var formats = make([]*formatXLV3, len(bootstrapDisks))
+	var formats = make([]*formatXLV3, len(storageDisks))
 
 	// Load format from each disk in parallel
-	for index, disk := range bootstrapDisks {
+	for index, disk := range storageDisks {
 		if disk == nil {
 			sErrs[index] = errDiskNotFound
 			continue
@@ -344,10 +335,9 @@ func loadFormatXLAll(endpoints EndpointList) ([]*formatXLV3, []error) {
 		// Launch go-routine per disk.
 		go func(index int, disk StorageAPI) {
 			defer wg.Done()
+
 			format, lErr := loadFormatXL(disk)
 			if lErr != nil {
-				// close the internal connection, to avoid fd leaks.
-				disk.Close()
 				sErrs[index] = lErr
 				return
 			}
@@ -530,12 +520,7 @@ func formatXLV3Check(reference *formatXLV3, format *formatXLV3) error {
 }
 
 // saveFormatXLAll - populates `format.json` on disks in its order.
-func saveFormatXLAll(endpoints EndpointList, formats []*formatXLV3) error {
-	storageDisks, err := initStorageDisks(endpoints)
-	if err != nil {
-		return err
-	}
-
+func saveFormatXLAll(storageDisks []StorageAPI, formats []*formatXLV3) error {
 	var errs = make([]error, len(storageDisks))
 
 	var wg = &sync.WaitGroup{}
@@ -556,17 +541,25 @@ func saveFormatXLAll(endpoints EndpointList, formats []*formatXLV3) error {
 	// Wait for the routines to finish.
 	wg.Wait()
 
-	writeQuorum := len(endpoints)/2 + 1
+	writeQuorum := len(storageDisks)/2 + 1
 	return reduceWriteQuorumErrs(errs, nil, writeQuorum)
 }
 
+// relinquishes the underlying connection for all storage disks.
+func closeStorageDisks(storageDisks []StorageAPI) {
+	for _, disk := range storageDisks {
+		if disk == nil {
+			continue
+		}
+		disk.Close()
+	}
+}
+
 // Initialize storage disks based on input arguments.
 func initStorageDisks(endpoints EndpointList) ([]StorageAPI, error) {
 	// Bootstrap disks.
 	storageDisks := make([]StorageAPI, len(endpoints))
 	for index, endpoint := range endpoints {
-		// Intentionally ignore disk not found errors. XL is designed
-		// to handle these errors internally.
 		storage, err := newStorageAPI(endpoint)
 		if err != nil && err != errDiskNotFound {
 			return nil, err
@@ -598,12 +591,7 @@ func formatXLV3ThisEmpty(formats []*formatXLV3) bool {
 }
 
 // fixFormatXLV3 - fix format XL configuration on all disks.
-func fixFormatXLV3(endpoints EndpointList, formats []*formatXLV3) error {
-	storageDisks, err := initStorageDisks(endpoints)
-	if err != nil {
-		return err
-	}
-
+func fixFormatXLV3(storageDisks []StorageAPI, endpoints EndpointList, formats []*formatXLV3) error {
 	for i, format := range formats {
 		if format == nil || !endpoints[i].IsLocal {
 			continue
@@ -617,7 +605,7 @@ func fixFormatXLV3(endpoints EndpointList, formats []*formatXLV3) error {
 		}
 		if format.XL.This == "" {
 			formats[i].XL.This = format.XL.Sets[0][i]
-			if err = saveFormatXL(storageDisks[i], formats[i]); err != nil {
+			if err := saveFormatXL(storageDisks[i], formats[i]); err != nil {
 				return err
 			}
 		}
@@ -626,9 +614,9 @@ func fixFormatXLV3(endpoints EndpointList, formats []*formatXLV3) error {
 }
 
 // initFormatXL - save XL format configuration on all disks.
-func initFormatXL(endpoints EndpointList, setCount, disksPerSet int) (format *formatXLV3, err error) {
+func initFormatXL(storageDisks []StorageAPI, setCount, disksPerSet int) (format *formatXLV3, err error) {
 	format = newFormatXLV3(setCount, disksPerSet)
-	formats := make([]*formatXLV3, len(endpoints))
+	formats := make([]*formatXLV3, len(storageDisks))
 
 	for i := 0; i < setCount; i++ {
 		for j := 0; j < disksPerSet; j++ {
@@ -639,12 +627,12 @@ func initFormatXL(endpoints EndpointList, setCount, disksPerSet int) (format *fo
 	}
 
 	// Initialize meta volume, if volume already exists ignores it.
-	if err = initFormatXLMetaVolume(endpoints, formats); err != nil {
+	if err = initFormatXLMetaVolume(storageDisks, formats); err != nil {
 		return format, fmt.Errorf("Unable to initialize '.minio.sys' meta volume, %s", err)
 	}
 
 	// Save formats `format.json` across all disks.
-	if err = saveFormatXLAll(endpoints, formats); err != nil {
+	if err = saveFormatXLAll(storageDisks, formats); err != nil {
 		return nil, err
 	}
 
@@ -675,12 +663,7 @@ func makeFormatXLMetaVolumes(disk StorageAPI) error {
 var initMetaVolIgnoredErrs = append(baseIgnoredErrs, errVolumeExists)
 
 // Initializes meta volume on all input storage disks.
-func initFormatXLMetaVolume(endpoints EndpointList, formats []*formatXLV3) error {
-	storageDisks, err := initStorageDisks(endpoints)
-	if err != nil {
-		return err
-	}
-
+func initFormatXLMetaVolume(storageDisks []StorageAPI, formats []*formatXLV3) error {
 	// This happens for the first time, but keep this here since this
 	// is the only place where it can be made expensive optimizing all
 	// other calls. Create minio meta volume, if it doesn't exist yet.

cmd/format-xl_test.go

@@ -87,6 +87,11 @@ func TestFixFormatV3(t *testing.T) {
 	}
 	endpoints := mustGetNewEndpointList(xlDirs...)
 
+	storageDisks, err := initStorageDisks(endpoints)
+	if err != nil {
+		t.Fatal(err)
+	}
+
 	format := newFormatXLV3(1, 8)
 	formats := make([]*formatXLV3, 8)
 
@@ -96,17 +101,18 @@ func TestFixFormatV3(t *testing.T) {
 		formats[j] = &newFormat
 	}
 
-	if err = initFormatXLMetaVolume(endpoints, formats); err != nil {
+	if err = initFormatXLMetaVolume(storageDisks, formats); err != nil {
 		t.Fatal(err)
 	}
 
 	formats[1] = nil
 	expThis := formats[2].XL.This
 	formats[2].XL.This = ""
-	if err := fixFormatXLV3(endpoints, formats); err != nil {
+	if err := fixFormatXLV3(storageDisks, endpoints, formats); err != nil {
 		t.Fatal(err)
 	}
-	newFormats, errs := loadFormatXLAll(endpoints)
+
+	newFormats, errs := loadFormatXLAll(storageDisks)
 	for _, err := range errs {
 		if err != nil && errors.Cause(err) != errUnformattedDisk {
 			t.Fatal(err)

cmd/prepare-storage.go

@@ -88,6 +88,95 @@ func formatXLCleanupTmpLocalEndpoints(endpoints EndpointList) error {
 	return nil
 }
 
+// validate reference format against list of XL formats.
+func validateXLFormats(format *formatXLV3, formats []*formatXLV3, endpoints EndpointList, setCount, drivesPerSet int) error {
+	for i := range formats {
+		if formats[i] == nil {
+			continue
+		}
+		if err := formatXLV3Check(format, formats[i]); err != nil {
+			return fmt.Errorf("%s format error: %s", endpoints[i], err)
+		}
+	}
+	if len(format.XL.Sets) != setCount {
+		return fmt.Errorf("Current backend format is inconsistent with input args (%s), Expected set count %d, got %d", endpoints, len(format.XL.Sets), setCount)
+	}
+	if len(format.XL.Sets[0]) != drivesPerSet {
+		return fmt.Errorf("Current backend format is inconsistent with input args (%s), Expected drive count per set %d, got %d", endpoints, len(format.XL.Sets[0]), drivesPerSet)
+	}
+	return nil
+}
+
+// Following error message is added to fix a regression in release
+// RELEASE.2018-03-16T22-52-12Z after migrating v1 to v2 to v3. This
+// migration failed to capture '.This' field properly which indicates
+// the disk UUID association. Below error message is returned when
+// we see this situation in format.json, for more info refer
+// https://github.com/minio/minio/issues/5667
+var errXLV3ThisEmpty = fmt.Errorf("XL format version 3 has This field empty")
+
+// connect to list of endpoints and load all XL disk formats, validate the formats are correct
+// and are in quorum, if no formats are found attempt to initialize all of them for the first
+// time. additionally make sure to close all the disks used in this attempt.
+func connectLoadInitFormats(firstDisk bool, endpoints EndpointList, setCount, drivesPerSet int) (*formatXLV3, error) {
+	storageDisks, err := initStorageDisks(endpoints)
+	if err != nil {
+		return nil, err
+	}
+	defer closeStorageDisks(storageDisks)
+
+	// Attempt to load all `format.json` from all disks.
+	formatConfigs, sErrs := loadFormatXLAll(storageDisks)
+
+	// Pre-emptively check if one of the formatted disks
+	// is invalid. This function returns success for the
+	// most part unless one of the formats is not consistent
+	// with expected XL format. For example if a user is
+	// trying to pool FS backend into an XL set.
+	if err = checkFormatXLValues(formatConfigs); err != nil {
+		return nil, err
+	}
+
+	for i, sErr := range sErrs {
+		if _, ok := formatCriticalErrors[errors.Cause(sErr)]; ok {
+			return nil, fmt.Errorf("Disk %s: %s", endpoints[i], sErr)
+		}
+	}
+
+	if shouldInitXLDisks(sErrs) {
+		if !firstDisk {
+			return nil, errNotFirstDisk
+		}
+		return initFormatXL(storageDisks, setCount, drivesPerSet)
+	}
+
+	// Following function is added to fix a regressions which was introduced
+	// in release RELEASE.2018-03-16T22-52-12Z after migrating v1 to v2 to v3.
+	// This migration failed to capture '.This' field properly which indicates
+	// the disk UUID association. Below function is called to handle and fix
+	// this regression, for more info refer https://github.com/minio/minio/issues/5667
+	if err = fixFormatXLV3(storageDisks, endpoints, formatConfigs); err != nil {
+		return nil, err
+	}
+
+	// If any of the .This field is still empty, we return error.
+	if formatXLV3ThisEmpty(formatConfigs) {
+		return nil, errXLV3ThisEmpty
+	}
+
+	format, err := getFormatXLInQuorum(formatConfigs)
+	if err != nil {
+		return nil, err
+	}
+
+	// Validate all format configs with reference format.
+	if err = validateXLFormats(format, formatConfigs, endpoints, setCount, drivesPerSet); err != nil {
+		return nil, err
+	}
+
+	return format, nil
+}
+
 // Format disks before initialization of object layer.
 func waitForFormatXL(firstDisk bool, endpoints EndpointList, setCount, disksPerSet int) (format *formatXLV3, err error) {
 	if len(endpoints) == 0 || setCount == 0 || disksPerSet == 0 {
@@ -121,65 +210,26 @@ func waitForFormatXL(firstDisk bool, endpoints EndpointList, setCount, disksPerS
 	for {
 		select {
 		case _ = <-retryTimerCh:
-			// Attempt to load all `format.json` from all disks.
-			formatConfigs, sErrs := loadFormatXLAll(endpoints)
-
-			// Pre-emptively check if one of the formatted disks
-			// is invalid. This function returns success for the
-			// most part unless one of the formats is not consistent
-			// with expected XL format. For example if a user is
-			// trying to pool FS backend into an XL set.
-			if err = checkFormatXLValues(formatConfigs); err != nil {
-				return nil, err
-			}
-
-			for i, sErr := range sErrs {
-				if _, ok := formatCriticalErrors[errors.Cause(sErr)]; ok {
-					return nil, fmt.Errorf("Disk %s: %s", endpoints[i], sErr)
-				}
-			}
-
-			if shouldInitXLDisks(sErrs) {
-				if !firstDisk {
+			format, err := connectLoadInitFormats(firstDisk, endpoints, setCount, disksPerSet)
+			if err != nil {
+				switch err {
+				case errNotFirstDisk:
+					// Fresh setup, wait for first server to be up.
 					console.Println("Waiting for the first server to format the disks.")
 					continue
+				case errXLReadQuorum:
+					// no quorum available continue to wait for minimum number of servers.
+					console.Printf("Waiting for a minimum of %d disks to come online (elapsed %s)\n", len(endpoints)/2, getElapsedTime())
+					continue
+				case errXLV3ThisEmpty:
+					// need to wait for this error to be healed, so continue.
+					continue
+				default:
+					// For all other unhandled errors we exit and fail.
+					return nil, err
 				}
-				return initFormatXL(endpoints, setCount, disksPerSet)
-			}
-
-			// Following function is added to fix a regressions which was introduced
-			// in release RELEASE.2018-03-16T22-52-12Z after migrating v1 to v2 to v3.
-			// This migration failed to capture '.This' field properly which indicates
-			// the disk UUID association. Below function is called to handle and fix
-			// this regression, for more info refer https://github.com/minio/minio/issues/5667
-			if err = fixFormatXLV3(endpoints, formatConfigs); err != nil {
-				return nil, err
-			}
-
-			// If any of the .This field is still empty we wait them to be fixed.
-			if formatXLV3ThisEmpty(formatConfigs) {
-				continue
-			}
-
-			format, err = getFormatXLInQuorum(formatConfigs)
-			if err == nil {
-				for i := range formatConfigs {
-					if formatConfigs[i] == nil {
-						continue
-					}
-					if err = formatXLV3Check(format, formatConfigs[i]); err != nil {
-						return nil, fmt.Errorf("%s format error: %s", endpoints[i], err)
-					}
-				}
-				if len(format.XL.Sets) != globalXLSetCount {
-					return nil, fmt.Errorf("Current backend format is inconsistent with input args (%s), Expected set count %d, got %d", endpoints, len(format.XL.Sets), globalXLSetCount)
-				}
-				if len(format.XL.Sets[0]) != globalXLSetDriveCount {
-					return nil, fmt.Errorf("Current backend format is inconsistent with input args (%s), Expected drive count per set %d, got %d", endpoints, len(format.XL.Sets[0]), globalXLSetDriveCount)
-				}
-				return format, nil
 			}
-			console.Printf("Waiting for a minimum of %d disks to come online (elapsed %s)\n", len(endpoints)/2, getElapsedTime())
+			return format, nil
 		case <-globalOSSignalCh:
 			return nil, fmt.Errorf("Initializing data volumes gracefully stopped")
 		}

cmd/typed-errors.go

@@ -58,3 +58,7 @@ var errInvalidRange = errors.New("Invalid range")
 // errInvalidRangeSource - returned when given range value exceeds
 // the source object size.
 var errInvalidRangeSource = errors.New("Range specified exceeds source object size")
+
+// error returned by disks which are to be initialized are waiting for the
+// first server to initialize them in distributed set to initialize them.
+var errNotFirstDisk = errors.New("Not first disk")

cmd/xl-sets.go

@@ -101,7 +101,7 @@ func connectEndpoint(endpoint Endpoint) (StorageAPI, *formatXLV3, error) {
 
 	format, err := loadFormatXL(disk)
 	if err != nil {
-		// close the internal connection, to avoid fd leaks.
+		// Close the internal connection to avoid connection leaks.
 		disk.Close()
 		return nil, nil, err
 	}
@@ -147,6 +147,8 @@ func (s *xlSets) connectDisks() {
 		i, j, err := findDiskIndex(s.format, format)
 		s.formatMu.RUnlock()
 		if err != nil {
+			// Close the internal connection to avoid connection leaks.
+			disk.Close()
 			printEndpointError(endpoint, err)
 			continue
 		}
@@ -215,19 +217,8 @@ func newXLSets(endpoints EndpointList, format *formatXLV3, setCount int, drivesP
 		go s.sets[i].cleanupStaleMultipartUploads(globalMultipartCleanupInterval, globalMultipartExpiry, globalServiceDoneCh)
 	}
 
-	for _, endpoint := range endpoints {
-		disk, nformat, err := connectEndpoint(endpoint)
-		if err != nil {
-			errorIf(err, "Unable to connect to endpoint %s", endpoint)
-			continue
-		}
-		i, j, err := findDiskIndex(format, nformat)
-		if err != nil {
-			errorIf(err, "Unable to find the endpoint %s in reference format", endpoint)
-			continue
-		}
-		s.xlDisks[i][j] = disk
-	}
+	// Connect disks right away.
+	s.connectDisks()
 
 	// Initialize and load bucket policies.
 	var err error
@@ -267,7 +258,19 @@ func (s *xlSets) StorageInfo(ctx context.Context) StorageInfo {
 	storageInfo.Backend.RRSCData = rrSCData
 	storageInfo.Backend.RRSCParity = rrSCparity
 
-	formats, sErrs := loadFormatXLAll(s.endpoints)
+	storageInfo.Backend.Sets = make([][]madmin.DriveInfo, s.setCount)
+	for i := range storageInfo.Backend.Sets {
+		storageInfo.Backend.Sets[i] = make([]madmin.DriveInfo, s.drivesPerSet)
+	}
+
+	storageDisks, err := initStorageDisks(s.endpoints)
+	if err != nil {
+		return storageInfo
+	}
+	defer closeStorageDisks(storageDisks)
+
+	formats, sErrs := loadFormatXLAll(storageDisks)
+
 	drivesInfo := formatsToDrivesInfo(s.endpoints, formats, sErrs)
 	refFormat, err := getFormatXLInQuorum(formats)
 	if err != nil {
@@ -276,11 +279,6 @@ func (s *xlSets) StorageInfo(ctx context.Context) StorageInfo {
 		return storageInfo
 	}
 
-	storageInfo.Backend.Sets = make([][]madmin.DriveInfo, s.setCount)
-	for i := range storageInfo.Backend.Sets {
-		storageInfo.Backend.Sets[i] = make([]madmin.DriveInfo, s.drivesPerSet)
-	}
-
 	// fill all the available/online endpoints
 	for _, drive := range drivesInfo {
 		if drive.UUID == "" {
@@ -921,8 +919,14 @@ func (s *xlSets) HealFormat(ctx context.Context, dryRun bool) (madmin.HealResult
 	}
 	defer formatLock.Unlock()
 
-	formats, sErrs := loadFormatXLAll(s.endpoints)
-	if err := checkFormatXLValues(formats); err != nil {
+	storageDisks, err := initStorageDisks(s.endpoints)
+	if err != nil {
+		return madmin.HealResultItem{}, err
+	}
+	defer closeStorageDisks(storageDisks)
+
+	formats, sErrs := loadFormatXLAll(storageDisks)
+	if err = checkFormatXLValues(formats); err != nil {
 		return madmin.HealResultItem{}, err
 	}
 
@@ -1025,12 +1029,12 @@ func (s *xlSets) HealFormat(ctx context.Context, dryRun bool) (madmin.HealResult
 
 		// Initialize meta volume, if volume already exists ignores it, all disks which
 		// are not found are ignored as well.
-		if err = initFormatXLMetaVolume(s.endpoints, tmpNewFormats); err != nil {
+		if err = initFormatXLMetaVolume(storageDisks, tmpNewFormats); err != nil {
 			return madmin.HealResultItem{}, fmt.Errorf("Unable to initialize '.minio.sys' meta volume, %s", err)
 		}
 
 		// Save formats `format.json` across all disks.
-		if err = saveFormatXLAll(s.endpoints, tmpNewFormats); err != nil {
+		if err = saveFormatXLAll(storageDisks, tmpNewFormats); err != nil {
 			return madmin.HealResultItem{}, err
 		}
 

cmd/xl-v1.go

@@ -59,14 +59,7 @@ var xlTreeWalkIgnoredErrs = append(baseIgnoredErrs, errDiskAccessDenied, errVolu
 // Shutdown function for object storage interface.
 func (xl xlObjects) Shutdown(ctx context.Context) error {
 	// Add any object layer shutdown activities here.
-	for _, disk := range xl.getDisks() {
-		// This closes storage rpc client connections if any.
-		// Otherwise this is a no-op.
-		if disk == nil {
-			continue
-		}
-		disk.Close()
-	}
+	closeStorageDisks(xl.getDisks())
 	return nil
 }