cache: offline uploading

backend/cache/cache_mount_unix_test.go

@@ -0,0 +1,78 @@
+// +build !plan9,!windows,go1.7
+
+package cache_test
+
+import (
+	"os"
+	"testing"
+	"time"
+
+	"bazil.org/fuse"
+	fusefs "bazil.org/fuse/fs"
+	"github.com/ncw/rclone/cmd/mount"
+	"github.com/ncw/rclone/cmd/mountlib"
+	"github.com/ncw/rclone/fs"
+	"github.com/stretchr/testify/require"
+)
+
+func (r *run) mountFs(t *testing.T, f fs.Fs) {
+	device := f.Name() + ":" + f.Root()
+	var options = []fuse.MountOption{
+		fuse.MaxReadahead(uint32(mountlib.MaxReadAhead)),
+		fuse.Subtype("rclone"),
+		fuse.FSName(device), fuse.VolumeName(device),
+		fuse.NoAppleDouble(),
+		fuse.NoAppleXattr(),
+		fuse.AllowOther(),
+	}
+	err := os.MkdirAll(r.mntDir, os.ModePerm)
+	require.NoError(t, err)
+	c, err := fuse.Mount(r.mntDir, options...)
+	require.NoError(t, err)
+	filesys := mount.NewFS(f)
+	server := fusefs.New(c, nil)
+
+	// Serve the mount point in the background returning error to errChan
+	r.unmountRes = make(chan error, 1)
+	go func() {
+		err := server.Serve(filesys)
+		closeErr := c.Close()
+		if err == nil {
+			err = closeErr
+		}
+		r.unmountRes <- err
+	}()
+
+	// check if the mount process has an error to report
+	<-c.Ready
+	require.NoError(t, c.MountError)
+
+	r.unmountFn = func() error {
+		// Shutdown the VFS
+		filesys.VFS.Shutdown()
+		return fuse.Unmount(r.mntDir)
+	}
+
+	r.vfs = filesys.VFS
+	r.isMounted = true
+}
+
+func (r *run) unmountFs(t *testing.T, f fs.Fs) {
+	var err error
+
+	for i := 0; i < 4; i++ {
+		err = r.unmountFn()
+		if err != nil {
+			//log.Printf("signal to umount failed - retrying: %v", err)
+			time.Sleep(3 * time.Second)
+			continue
+		}
+		break
+	}
+	require.NoError(t, err)
+	err = <-r.unmountRes
+	require.NoError(t, err)
+	err = r.vfs.CleanUp()
+	require.NoError(t, err)
+	r.isMounted = false
+}

backend/cache/cache_mount_windows_test.go

@@ -0,0 +1,124 @@
+// +build windows,go1.7
+
+package cache_test
+
+import (
+	"fmt"
+	"os"
+	"testing"
+	"time"
+
+	"github.com/billziss-gh/cgofuse/fuse"
+	"github.com/ncw/rclone/cmd/cmount"
+	"github.com/ncw/rclone/cmd/mountlib"
+	"github.com/ncw/rclone/fs"
+	"github.com/pkg/errors"
+	"github.com/stretchr/testify/require"
+)
+
+// waitFor runs fn() until it returns true or the timeout expires
+func waitFor(fn func() bool) (ok bool) {
+	const totalWait = 10 * time.Second
+	const individualWait = 10 * time.Millisecond
+	for i := 0; i < int(totalWait/individualWait); i++ {
+		ok = fn()
+		if ok {
+			return ok
+		}
+		time.Sleep(individualWait)
+	}
+	return false
+}
+
+func (r *run) mountFs(t *testing.T, f fs.Fs) {
+	// FIXME implement cmount
+	t.Skip("windows not supported yet")
+
+	device := f.Name() + ":" + f.Root()
+	options := []string{
+		"-o", "fsname=" + device,
+		"-o", "subtype=rclone",
+		"-o", fmt.Sprintf("max_readahead=%d", mountlib.MaxReadAhead),
+		"-o", "uid=-1",
+		"-o", "gid=-1",
+		"-o", "allow_other",
+		// This causes FUSE to supply O_TRUNC with the Open
+		// call which is more efficient for cmount.  However
+		// it does not work with cgofuse on Windows with
+		// WinFSP so cmount must work with or without it.
+		"-o", "atomic_o_trunc",
+		"--FileSystemName=rclone",
+	}
+
+	fsys := cmount.NewFS(f)
+	host := fuse.NewFileSystemHost(fsys)
+
+	// Serve the mount point in the background returning error to errChan
+	r.unmountRes = make(chan error, 1)
+	go func() {
+		var err error
+		ok := host.Mount(r.mntDir, options)
+		if !ok {
+			err = errors.New("mount failed")
+		}
+		r.unmountRes <- err
+	}()
+
+	// unmount
+	r.unmountFn = func() error {
+		// Shutdown the VFS
+		fsys.VFS.Shutdown()
+		if host.Unmount() {
+			if !waitFor(func() bool {
+				_, err := os.Stat(r.mntDir)
+				return err != nil
+			}) {
+				t.Fatalf("mountpoint %q didn't disappear after unmount - continuing anyway", r.mntDir)
+			}
+			return nil
+		}
+		return errors.New("host unmount failed")
+	}
+
+	// Wait for the filesystem to become ready, checking the file
+	// system didn't blow up before starting
+	select {
+	case err := <-r.unmountRes:
+		require.NoError(t, err)
+	case <-time.After(time.Second * 3):
+	}
+
+	// Wait for the mount point to be available on Windows
+	// On Windows the Init signal comes slightly before the mount is ready
+	if !waitFor(func() bool {
+		_, err := os.Stat(r.mntDir)
+		return err == nil
+	}) {
+		t.Errorf("mountpoint %q didn't became available on mount", r.mntDir)
+	}
+
+	r.vfs = fsys.VFS
+	r.isMounted = true
+}
+
+func (r *run) unmountFs(t *testing.T, f fs.Fs) {
+	// FIXME implement cmount
+	t.Skip("windows not supported yet")
+	var err error
+
+	for i := 0; i < 4; i++ {
+		err = r.unmountFn()
+		if err != nil {
+			//log.Printf("signal to umount failed - retrying: %v", err)
+			time.Sleep(3 * time.Second)
+			continue
+		}
+		break
+	}
+	require.NoError(t, err)
+	err = <-r.unmountRes
+	require.NoError(t, err)
+	err = r.vfs.CleanUp()
+	require.NoError(t, err)
+	r.isMounted = false
+}

backend/cache/directory.go

@@ -93,15 +93,7 @@ func (d *Directory) String() string {
 
 // Remote returns the remote path
 func (d *Directory) Remote() string {
-	p := cleanPath(path.Join(d.Dir, d.Name))
+	return d.CacheFs.cleanRootFromPath(d.abs())
-	if d.CacheFs.Root() != "" {
-		p = p[len(d.CacheFs.Root()):] // trim out root
-		if len(p) > 0 {               // remove first separator
-			p = p[1:]
-		}
-	}
-
-	return p
 }
 
 // abs returns the absolute path to the dir

backend/cache/handle.go

@@ -9,14 +9,42 @@ import (
 	"sync"
 	"time"
 
+	"path"
+	"runtime"
+	"strings"
+
 	"github.com/ncw/rclone/fs"
+	"github.com/ncw/rclone/fs/operations"
 	"github.com/pkg/errors"
 )
 
+var uploaderMap = make(map[string]*backgroundWriter)
+var uploaderMapMx sync.Mutex
+
+// initBackgroundUploader returns a single instance
+func initBackgroundUploader(fs *Fs) (*backgroundWriter, error) {
+	// write lock to create one
+	uploaderMapMx.Lock()
+	defer uploaderMapMx.Unlock()
+	if b, ok := uploaderMap[fs.String()]; ok {
+		// if it was already started we close it so that it can be started again
+		if b.running {
+			b.close()
+		} else {
+			return b, nil
+		}
+	}
+
+	bb := newBackgroundWriter(fs)
+	uploaderMap[fs.String()] = bb
+	return uploaderMap[fs.String()], nil
+}
+
 // Handle is managing the read/write/seek operations on an open handle
 type Handle struct {
 	cachedObject   *Object
-	memory         ChunkStorage
+	cfs            *Fs
+	memory         *Memory
 	preloadQueue   chan int64
 	preloadOffset  int64
 	offset         int64
@@ -31,20 +59,21 @@ type Handle struct {
 }
 
 // NewObjectHandle returns a new Handle for an existing Object
-func NewObjectHandle(o *Object) *Handle {
+func NewObjectHandle(o *Object, cfs *Fs) *Handle {
 	r := &Handle{
 		cachedObject:  o,
+		cfs:           cfs,
 		offset:        0,
 		preloadOffset: -1, // -1 to trigger the first preload
 
-		UseMemory: o.CacheFs.chunkMemory,
+		UseMemory: cfs.chunkMemory,
 		reading:   false,
 	}
 	r.seenOffsets = make(map[int64]bool)
 	r.memory = NewMemory(-1)
 
 	// create a larger buffer to queue up requests
-	r.preloadQueue = make(chan int64, o.CacheFs.totalWorkers*10)
+	r.preloadQueue = make(chan int64, r.cfs.totalWorkers*10)
 	r.confirmReading = make(chan bool)
 	r.startReadWorkers()
 	return r
@@ -52,11 +81,11 @@ func NewObjectHandle(o *Object) *Handle {
 
 // cacheFs is a convenience method to get the parent cache FS of the object's manager
 func (r *Handle) cacheFs() *Fs {
-	return r.cachedObject.CacheFs
+	return r.cfs
 }
 
 // storage is a convenience method to get the persistent storage of the object's manager
-func (r *Handle) storage() Storage {
+func (r *Handle) storage() *Persistent {
 	return r.cacheFs().cache
 }
 
@@ -76,7 +105,7 @@ func (r *Handle) startReadWorkers() {
 		if !r.cacheFs().plexConnector.isConnected() {
 			err := r.cacheFs().plexConnector.authenticate()
 			if err != nil {
-				fs.Infof(r, "failed to authenticate to Plex: %v", err)
+				fs.Errorf(r, "failed to authenticate to Plex: %v", err)
 			}
 		}
 		if r.cacheFs().plexConnector.isConnected() {
@@ -113,7 +142,7 @@ func (r *Handle) scaleWorkers(desired int) {
 	}
 	// ignore first scale out from 0
 	if current != 0 {
-		fs.Infof(r, "scale workers to %v", desired)
+		fs.Debugf(r, "scale workers to %v", desired)
 	}
 }
 
@@ -156,7 +185,6 @@ func (r *Handle) queueOffset(offset int64) {
 		if r.UseMemory {
 			go r.memory.CleanChunksByNeed(offset)
 		}
-		go r.cacheFs().CleanUpCache(false)
 		r.confirmExternalReading()
 		r.preloadOffset = offset
 
@@ -305,7 +333,6 @@ func (r *Handle) Close() error {
 		}
 	}
 
-	go r.cacheFs().CleanUpCache(false)
 	fs.Debugf(r, "cache reader closed %v", r.offset)
 	return nil
 }
@@ -357,11 +384,11 @@ func (w *worker) String() string {
 //   - if it supports seeking it will seek to the desired offset and return the same reader
 //   - if it doesn't support seeking it will close a possible existing one and open at the desired offset
 //   - if there's no reader associated with this worker, it will create one
-func (w *worker) reader(offset, end int64) (io.ReadCloser, error) {
+func (w *worker) reader(offset, end int64, closeOpen bool) (io.ReadCloser, error) {
 	var err error
 	r := w.rc
 	if w.rc == nil {
-		r, err = w.r.cacheFs().OpenRateLimited(func() (io.ReadCloser, error) {
+		r, err = w.r.cacheFs().openRateLimited(func() (io.ReadCloser, error) {
 			return w.r.cachedObject.Object.Open(&fs.SeekOption{Offset: offset}, &fs.RangeOption{Start: offset, End: end})
 		})
 		if err != nil {
@@ -370,14 +397,16 @@ func (w *worker) reader(offset, end int64) (io.ReadCloser, error) {
 		return r, nil
 	}
 
+	if !closeOpen {
 		seekerObj, ok := r.(io.Seeker)
 		if ok {
 			_, err = seekerObj.Seek(offset, os.SEEK_SET)
 			return r, err
 		}
+	}
 
 	_ = w.rc.Close()
-	return w.r.cacheFs().OpenRateLimited(func() (io.ReadCloser, error) {
+	return w.r.cacheFs().openRateLimited(func() (io.ReadCloser, error) {
 		r, err = w.r.cachedObject.Object.Open(&fs.SeekOption{Offset: offset}, &fs.RangeOption{Start: offset, End: end})
 		if err != nil {
 			return nil, err
@@ -463,10 +492,18 @@ func (w *worker) download(chunkStart, chunkEnd int64, retry int) {
 		time.Sleep(time.Second * time.Duration(retry))
 	}
 
-	w.rc, err = w.reader(chunkStart, chunkEnd)
+	closeOpen := false
+	if retry > 0 {
+		closeOpen = true
+	}
+	w.rc, err = w.reader(chunkStart, chunkEnd, closeOpen)
 	// we seem to be getting only errors so we abort
 	if err != nil {
 		fs.Errorf(w, "object open failed %v: %v", chunkStart, err)
+		err = w.r.cachedObject.refreshFromSource(true)
+		if err != nil {
+			fs.Errorf(w, "%v", err)
+		}
 		w.download(chunkStart, chunkEnd, retry+1)
 		return
 	}
@@ -476,6 +513,10 @@ func (w *worker) download(chunkStart, chunkEnd int64, retry int) {
 	sourceRead, err = io.ReadFull(w.rc, data)
 	if err != nil && err != io.EOF && err != io.ErrUnexpectedEOF {
 		fs.Errorf(w, "failed to read chunk %v: %v", chunkStart, err)
+		err = w.r.cachedObject.refreshFromSource(true)
+		if err != nil {
+			fs.Errorf(w, "%v", err)
+		}
 		w.download(chunkStart, chunkEnd, retry+1)
 		return
 	}
@@ -483,7 +524,7 @@ func (w *worker) download(chunkStart, chunkEnd int64, retry int) {
 	if err == io.ErrUnexpectedEOF {
 		fs.Debugf(w, "partial downloaded chunk %v", fs.SizeSuffix(chunkStart))
 	} else {
-		fs.Debugf(w, "downloaded chunk %v", fs.SizeSuffix(chunkStart))
+		fs.Debugf(w, "downloaded chunk %v", chunkStart)
 	}
 
 	if w.r.UseMemory {
@@ -499,6 +540,115 @@ func (w *worker) download(chunkStart, chunkEnd int64, retry int) {
 	}
 }
 
+const (
+	// BackgroundUploadStarted is a state for a temp file that has started upload
+	BackgroundUploadStarted = iota
+	// BackgroundUploadCompleted is a state for a temp file that has completed upload
+	BackgroundUploadCompleted
+	// BackgroundUploadError is a state for a temp file that has an error upload
+	BackgroundUploadError
+)
+
+// BackgroundUploadState is an entity that maps to an existing file which is stored on the temp fs
+type BackgroundUploadState struct {
+	Remote string
+	Status int
+	Error  error
+}
+
+type backgroundWriter struct {
+	fs       *Fs
+	stateCh  chan int
+	running  bool
+	notifyCh chan BackgroundUploadState
+}
+
+func newBackgroundWriter(f *Fs) *backgroundWriter {
+	b := &backgroundWriter{
+		fs:       f,
+		stateCh:  make(chan int),
+		notifyCh: make(chan BackgroundUploadState),
+	}
+
+	return b
+}
+
+func (b *backgroundWriter) close() {
+	b.stateCh <- 2
+}
+
+func (b *backgroundWriter) pause() {
+	b.stateCh <- 1
+}
+
+func (b *backgroundWriter) play() {
+	b.stateCh <- 0
+}
+
+func (b *backgroundWriter) notify(remote string, status int, err error) {
+	state := BackgroundUploadState{
+		Remote: remote,
+		Status: status,
+		Error:  err,
+	}
+	select {
+	case b.notifyCh <- state:
+		fs.Debugf(remote, "notified background upload state: %v", state.Status)
+	default:
+	}
+}
+
+func (b *backgroundWriter) run() {
+	state := 0
+	for {
+		b.running = true
+		select {
+		case s := <-b.stateCh:
+			state = s
+		default:
+			//
+		}
+		switch state {
+		case 1:
+			runtime.Gosched()
+			time.Sleep(time.Millisecond * 500)
+			continue
+		case 2:
+			b.running = false
+			return
+		}
+
+		absPath, err := b.fs.cache.getPendingUpload(b.fs.Root(), b.fs.tempWriteWait)
+		if err != nil || absPath == "" || !b.fs.isRootInPath(absPath) {
+			time.Sleep(time.Second)
+			continue
+		}
+
+		remote := b.fs.cleanRootFromPath(absPath)
+		b.notify(remote, BackgroundUploadStarted, nil)
+		fs.Infof(remote, "background upload: started upload")
+		err = operations.MoveFile(b.fs.UnWrap(), b.fs.tempFs, remote, remote)
+		if err != nil {
+			b.notify(remote, BackgroundUploadError, err)
+			_ = b.fs.cache.rollbackPendingUpload(absPath)
+			fs.Errorf(remote, "background upload: %v", err)
+			continue
+		}
+		fs.Infof(remote, "background upload: uploaded entry")
+		err = b.fs.cache.removePendingUpload(absPath)
+		if err != nil && !strings.Contains(err.Error(), "pending upload not found") {
+			fs.Errorf(remote, "background upload: %v", err)
+		}
+		parentCd := NewDirectory(b.fs, cleanPath(path.Dir(remote)))
+		err = b.fs.cache.ExpireDir(parentCd)
+		if err != nil {
+			fs.Errorf(parentCd, "background upload: cache expire error: %v", err)
+		}
+		fs.Infof(remote, "finished background upload")
+		b.notify(remote, BackgroundUploadCompleted, nil)
+	}
+}
+
 // Check the interfaces are satisfied
 var (
 	_ io.ReadCloser = (*Handle)(nil)

backend/cache/object.go

@@ -3,24 +3,28 @@
 package cache
 
 import (
-	"encoding/json"
 	"io"
 	"os"
 	"path"
 	"sync"
 	"time"
 
-	"strconv"
-
 	"github.com/ncw/rclone/fs"
 	"github.com/ncw/rclone/fs/hash"
 	"github.com/ncw/rclone/lib/readers"
+	"github.com/pkg/errors"
+)
+
+const (
+	objectInCache       = "Object"
+	objectPendingUpload = "TempObject"
 )
 
 // Object is a generic file like object that stores basic information about it
 type Object struct {
 	fs.Object `json:"-"`
 
+	ParentFs      fs.Fs                `json:"-"`        // parent fs
 	CacheFs       *Fs                  `json:"-"`        // cache fs
 	Name          string               `json:"name"`     // name of the directory
 	Dir           string               `json:"dir"`      // abs path of the object
@@ -29,79 +33,64 @@ type Object struct {
 	CacheStorable bool                 `json:"storable"` // says whether this object can be stored
 	CacheType     string               `json:"cacheType"`
 	CacheTs       time.Time            `json:"cacheTs"`
-	cacheHashes   map[hash.Type]string // all supported hashes cached
+	CacheHashes   map[hash.Type]string // all supported hashes cached
 
 	refreshMutex sync.Mutex
 }
 
 // NewObject builds one from a generic fs.Object
-func NewObject(f *Fs, remote string) *Object { //0745 379 768
+func NewObject(f *Fs, remote string) *Object {
 	fullRemote := path.Join(f.Root(), remote)
 	dir, name := path.Split(fullRemote)
 
+	cacheType := objectInCache
+	parentFs := f.UnWrap()
+	if f.tempWritePath != "" {
+		_, err := f.cache.SearchPendingUpload(fullRemote)
+		if err == nil { // queued for upload
+			cacheType = objectPendingUpload
+			parentFs = f.tempFs
+			fs.Debugf(fullRemote, "pending upload found")
+		}
+	}
+
 	co := &Object{
+		ParentFs:      parentFs,
 		CacheFs:       f,
 		Name:          cleanPath(name),
 		Dir:           cleanPath(dir),
 		CacheModTime:  time.Now().UnixNano(),
 		CacheSize:     0,
 		CacheStorable: false,
-		CacheType:     "Object",
+		CacheType:     cacheType,
 		CacheTs:       time.Now(),
 	}
 	return co
 }
 
-// MarshalJSON is needed to override the hashes map (needed to support older versions of Go)
-func (o *Object) MarshalJSON() ([]byte, error) {
-	hashes := make(map[string]string)
-	for k, v := range o.cacheHashes {
-		hashes[strconv.Itoa(int(k))] = v
-	}
-
-	type Alias Object
-	return json.Marshal(&struct {
-		Hashes map[string]string `json:"hashes"`
-		*Alias
-	}{
-		Alias:  (*Alias)(o),
-		Hashes: hashes,
-	})
-}
-
-// UnmarshalJSON is needed to override the CacheHashes map (needed to support older versions of Go)
-func (o *Object) UnmarshalJSON(b []byte) error {
-	type Alias Object
-	aux := &struct {
-		Hashes map[string]string `json:"hashes"`
-		*Alias
-	}{
-		Alias: (*Alias)(o),
-	}
-	if err := json.Unmarshal(b, &aux); err != nil {
-		return err
-	}
-
-	o.cacheHashes = make(map[hash.Type]string)
-	for k, v := range aux.Hashes {
-		ht, _ := strconv.Atoi(k)
-		o.cacheHashes[hash.Type(ht)] = v
-	}
-
-	return nil
-}
-
 // ObjectFromOriginal builds one from a generic fs.Object
 func ObjectFromOriginal(f *Fs, o fs.Object) *Object {
 	var co *Object
 	fullRemote := cleanPath(path.Join(f.Root(), o.Remote()))
-
 	dir, name := path.Split(fullRemote)
+
+	cacheType := objectInCache
+	parentFs := f.UnWrap()
+	if f.tempWritePath != "" {
+		_, err := f.cache.SearchPendingUpload(fullRemote)
+		if err == nil { // queued for upload
+			cacheType = objectPendingUpload
+			parentFs = f.tempFs
+			fs.Debugf(fullRemote, "pending upload found")
+		}
+	}
+
 	co = &Object{
+		ParentFs:  parentFs,
 		CacheFs:   f,
 		Name:      cleanPath(name),
 		Dir:       cleanPath(dir),
-		CacheType: "Object",
+		CacheType: cacheType,
 		CacheTs:   time.Now(),
 	}
 	co.updateData(o)
@@ -114,7 +103,7 @@ func (o *Object) updateData(source fs.Object) {
 	o.CacheSize = source.Size()
 	o.CacheStorable = source.Storable()
 	o.CacheTs = time.Now()
-	o.cacheHashes = make(map[hash.Type]string)
+	o.CacheHashes = make(map[hash.Type]string)
 }
 
 // Fs returns its FS info
@@ -133,14 +122,7 @@ func (o *Object) String() string {
 // Remote returns the remote path
 func (o *Object) Remote() string {
 	p := path.Join(o.Dir, o.Name)
-	if o.CacheFs.Root() != "" {
+	return o.CacheFs.cleanRootFromPath(p)
-		p = p[len(o.CacheFs.Root()):] // trim out root
-		if len(p) > 0 {               // remove first separator
-			p = p[1:]
-		}
-	}
-
-	return p
 }
 
 // abs returns the absolute path to the object
@@ -148,17 +130,6 @@ func (o *Object) abs() string {
 	return path.Join(o.Dir, o.Name)
 }
 
-// parentRemote returns the absolute path parent remote
-func (o *Object) parentRemote() string {
-	absPath := o.abs()
-	return cleanPath(path.Dir(absPath))
-}
-
-// parentDir returns the absolute path parent remote
-func (o *Object) parentDir() *Directory {
-	return NewDirectory(o.CacheFs, cleanPath(path.Dir(o.Remote())))
-}
-
 // ModTime returns the cached ModTime
 func (o *Object) ModTime() time.Time {
 	return time.Unix(0, o.CacheModTime)
@@ -175,17 +146,24 @@ func (o *Object) Storable() bool {
 }
 
 // refreshFromSource requests the original FS for the object in case it comes from a cached entry
-func (o *Object) refreshFromSource() error {
+func (o *Object) refreshFromSource(force bool) error {
 	o.refreshMutex.Lock()
 	defer o.refreshMutex.Unlock()
+	var err error
+	var liveObject fs.Object
 
-	if o.Object != nil {
+	if o.Object != nil && !force {
 		return nil
 	}
-
+	if o.isTempFile() {
-	liveObject, err := o.CacheFs.Fs.NewObject(o.Remote())
+		liveObject, err = o.ParentFs.NewObject(o.Remote())
+		err = errors.Wrapf(err, "in parent fs %v", o.ParentFs)
+	} else {
+		liveObject, err = o.CacheFs.Fs.NewObject(o.Remote())
+		err = errors.Wrapf(err, "in cache fs %v", o.CacheFs.Fs)
+	}
 	if err != nil {
-		fs.Errorf(o, "error refreshing object: %v", err)
+		fs.Errorf(o, "error refreshing object in : %v", err)
 		return err
 	}
 	o.updateData(liveObject)
@@ -196,7 +174,7 @@ func (o *Object) refreshFromSource() error {
 
 // SetModTime sets the ModTime of this object
 func (o *Object) SetModTime(t time.Time) error {
-	if err := o.refreshFromSource(); err != nil {
+	if err := o.refreshFromSource(false); err != nil {
 		return err
 	}
 
@@ -207,19 +185,19 @@ func (o *Object) SetModTime(t time.Time) error {
 
 	o.CacheModTime = t.UnixNano()
 	o.persist()
-	fs.Debugf(o.Fs(), "updated ModTime %v: %v", o, t)
+	fs.Debugf(o, "updated ModTime: %v", t)
 
 	return nil
 }
 
 // Open is used to request a specific part of the file using fs.RangeOption
 func (o *Object) Open(options ...fs.OpenOption) (io.ReadCloser, error) {
-	if err := o.refreshFromSource(); err != nil {
+	if err := o.refreshFromSource(true); err != nil {
 		return nil, err
 	}
 
 	var err error
-	cacheReader := NewObjectHandle(o)
+	cacheReader := NewObjectHandle(o, o.CacheFs)
 	var offset, limit int64 = 0, -1
 	for _, option := range options {
 		switch x := option.(type) {
@@ -239,23 +217,34 @@ func (o *Object) Open(options ...fs.OpenOption) (io.ReadCloser, error) {
 
 // Update will change the object data
 func (o *Object) Update(in io.Reader, src fs.ObjectInfo, options ...fs.OpenOption) error {
-	if err := o.refreshFromSource(); err != nil {
+	if err := o.refreshFromSource(false); err != nil {
 		return err
 	}
-	fs.Infof(o, "updating object contents with size %v", src.Size())
+	// pause background uploads if active
-
+	if o.CacheFs.tempWritePath != "" {
-	// deleting cached chunks and info to be replaced with new ones
+		o.CacheFs.backgroundRunner.pause()
-	_ = o.CacheFs.cache.RemoveObject(o.abs())
+		defer o.CacheFs.backgroundRunner.play()
+		// don't allow started uploads
+		if o.isTempFile() && o.tempFileStartedUpload() {
+			return errors.Errorf("%v is currently uploading, can't update", o)
+		}
+	}
+	fs.Debugf(o, "updating object contents with size %v", src.Size())
 
+	// FIXME use reliable upload
 	err := o.Object.Update(in, src, options...)
 	if err != nil {
 		fs.Errorf(o, "error updating source: %v", err)
 		return err
 	}
 
+	// deleting cached chunks and info to be replaced with new ones
+	_ = o.CacheFs.cache.RemoveObject(o.abs())
+
 	o.CacheModTime = src.ModTime().UnixNano()
 	o.CacheSize = src.Size()
-	o.cacheHashes = make(map[hash.Type]string)
+	o.CacheHashes = make(map[hash.Type]string)
+	o.CacheTs = time.Now()
 	o.persist()
 
 	return nil
@@ -263,41 +252,50 @@ func (o *Object) Update(in io.Reader, src fs.ObjectInfo, options ...fs.OpenOptio
 
 // Remove deletes the object from both the cache and the source
 func (o *Object) Remove() error {
-	if err := o.refreshFromSource(); err != nil {
+	if err := o.refreshFromSource(false); err != nil {
 		return err
 	}
+	// pause background uploads if active
+	if o.CacheFs.tempWritePath != "" {
+		o.CacheFs.backgroundRunner.pause()
+		defer o.CacheFs.backgroundRunner.play()
+		// don't allow started uploads
+		if o.isTempFile() && o.tempFileStartedUpload() {
+			return errors.Errorf("%v is currently uploading, can't delete", o)
+		}
+	}
 	err := o.Object.Remove()
 	if err != nil {
 		return err
 	}
-	fs.Infof(o, "removing object")
 
+	fs.Debugf(o, "removing object")
 	_ = o.CacheFs.cache.RemoveObject(o.abs())
-	return err
+	_ = o.CacheFs.cache.removePendingUpload(o.abs())
+	_ = o.CacheFs.cache.ExpireDir(NewDirectory(o.CacheFs, cleanPath(path.Dir(o.Remote()))))
+
+	return nil
 }
 
 // Hash requests a hash of the object and stores in the cache
 // since it might or might not be called, this is lazy loaded
 func (o *Object) Hash(ht hash.Type) (string, error) {
-	if o.cacheHashes == nil {
+	if o.CacheHashes == nil {
-		o.cacheHashes = make(map[hash.Type]string)
+		o.CacheHashes = make(map[hash.Type]string)
 	}
 
-	cachedHash, found := o.cacheHashes[ht]
+	cachedHash, found := o.CacheHashes[ht]
 	if found {
 		return cachedHash, nil
 	}
-
+	if err := o.refreshFromSource(false); err != nil {
-	if err := o.refreshFromSource(); err != nil {
 		return "", err
 	}
-
 	liveHash, err := o.Object.Hash(ht)
 	if err != nil {
 		return "", err
 	}
-
+	o.CacheHashes[ht] = liveHash
-	o.cacheHashes[ht] = liveHash
 
 	o.persist()
 	fs.Debugf(o, "object hash cached: %v", liveHash)
@@ -314,6 +312,25 @@ func (o *Object) persist() *Object {
 	return o
 }
 
+func (o *Object) isTempFile() bool {
+	_, err := o.CacheFs.cache.SearchPendingUpload(o.abs())
+	if err != nil {
+		o.CacheType = objectInCache
+		return false
+	}
+
+	o.CacheType = objectPendingUpload
+	return true
+}
+
+func (o *Object) tempFileStartedUpload() bool {
+	started, err := o.CacheFs.cache.SearchPendingUpload(o.abs())
+	if err != nil {
+		return false
+	}
+	return started
+}
+
 var (
 	_ fs.Object = (*Object)(nil)
 )

backend/cache/plex.go

@@ -12,6 +12,9 @@ import (
 
 	"sync"
 
+	"bytes"
+	"io/ioutil"
+
 	"github.com/ncw/rclone/fs"
 	"github.com/ncw/rclone/fs/config"
 )
@@ -127,6 +130,17 @@ func (p *plexConnector) isConfigured() bool {
 }
 
 func (p *plexConnector) isPlaying(co *Object) bool {
+	var err error
+
+	remote := co.Remote()
+	if cr, yes := p.f.isWrappedByCrypt(); yes {
+		remote, err = cr.DecryptFileName(co.Remote())
+		if err != nil {
+			fs.Errorf("plex", "can not decrypt wrapped file: %v", err)
+			return false
+		}
+	}
+
 	isPlaying := false
 	req, err := http.NewRequest("GET", fmt.Sprintf("%s/status/sessions", p.url.String()), nil)
 	if err != nil {
@@ -180,31 +194,12 @@ func (p *plexConnector) isPlaying(co *Object) bool {
 		if err != nil {
 			return false
 		}
-		var data map[string]interface{}
+		var data []byte
-		err = json.NewDecoder(resp.Body).Decode(&data)
+		data, err = ioutil.ReadAll(resp.Body)
 		if err != nil {
 			return false
 		}
-
+		if bytes.Contains(data, []byte(remote)) {
-		remote := co.Remote()
-		if cr, yes := co.CacheFs.isWrappedByCrypt(); yes {
-			remote, err = cr.DecryptFileName(co.Remote())
-			if err != nil {
-				fs.Errorf("plex", "can not decrypt wrapped file: %v", err)
-				continue
-			}
-		}
-		fpGen, ok := get(data, "MediaContainer", "Metadata", 0, "Media", 0, "Part", 0, "file")
-		if !ok {
-			fs.Errorf("plex", "failed to understand: %v", data)
-			continue
-		}
-		fp, ok := fpGen.(string)
-		if !ok {
-			fs.Errorf("plex", "failed to understand: %v", fp)
-			continue
-		}
-		if strings.Contains(fp, remote) {
 			isPlaying = true
 			break
 		}

backend/cache/storage_memory.go

@@ -14,8 +14,6 @@ import (
 
 // Memory is a wrapper of transient storage for a go-cache store
 type Memory struct {
-	ChunkStorage
-
 	db *cache.Cache
 }
 

backend/cache/storage_persistent.go

@@ -16,8 +16,11 @@ import (
 
 	"io/ioutil"
 
+	"fmt"
+
 	bolt "github.com/coreos/bbolt"
 	"github.com/ncw/rclone/fs"
+	"github.com/ncw/rclone/fs/walk"
 	"github.com/pkg/errors"
 )
 
@@ -26,6 +29,7 @@ const (
 	RootBucket   = "root"
 	RootTsBucket = "rootTs"
 	DataTsBucket = "dataTs"
+	tempBucket   = "pending"
 )
 
 // Features flags for this storage type
@@ -42,6 +46,12 @@ func GetPersistent(dbPath, chunkPath string, f *Features) (*Persistent, error) {
 	boltMapMx.Lock()
 	defer boltMapMx.Unlock()
 	if b, ok := boltMap[dbPath]; ok {
+		if !b.open {
+			err := b.connect()
+			if err != nil {
+				return nil, err
+			}
+		}
 		return b, nil
 	}
 
@@ -59,14 +69,25 @@ type chunkInfo struct {
 	Size   int64
 }
 
+type tempUploadInfo struct {
+	DestPath string
+	AddedOn  time.Time
+	Started  bool
+}
+
+// String representation of a tempUploadInfo
+func (t *tempUploadInfo) String() string {
+	return fmt.Sprintf("%v - %v (%v)", t.DestPath, t.Started, t.AddedOn)
+}
+
 // Persistent is a wrapper of persistent storage for a bolt.DB file
 type Persistent struct {
-	Storage
-
 	dbPath       string
 	dataPath     string
+	open         bool
 	db           *bolt.DB
 	cleanupMux   sync.Mutex
+	tempQueueMux sync.Mutex
 	features     *Features
 }
 
@@ -78,7 +99,7 @@ func newPersistent(dbPath, chunkPath string, f *Features) (*Persistent, error) {
 		features: f,
 	}
 
-	err := b.Connect()
+	err := b.connect()
 	if err != nil {
 		fs.Errorf(dbPath, "Error opening storage cache. Is there another rclone running on the same remote? %v", err)
 		return nil, err
@@ -92,41 +113,32 @@ func (b *Persistent) String() string {
 	return "<Cache DB> " + b.dbPath
 }
 
-// Connect creates a connection to the configured file
+// connect creates a connection to the configured file
 // refreshDb will delete the file before to create an empty DB if it's set to true
-func (b *Persistent) Connect() error {
+func (b *Persistent) connect() error {
-	var db *bolt.DB
 	var err error
 
-	if b.features.PurgeDb {
-		err := os.Remove(b.dbPath)
-		if err != nil {
-			fs.Errorf(b, "failed to remove cache file: %v", err)
-		}
-		err = os.RemoveAll(b.dataPath)
-		if err != nil {
-			fs.Errorf(b, "failed to remove cache data: %v", err)
-		}
-	}
-
 	err = os.MkdirAll(b.dataPath, os.ModePerm)
 	if err != nil {
 		return errors.Wrapf(err, "failed to create a data directory %q", b.dataPath)
 	}
-	db, err = bolt.Open(b.dbPath, 0644, &bolt.Options{Timeout: 1 * time.Second})
+	b.db, err = bolt.Open(b.dbPath, 0644, &bolt.Options{Timeout: 1 * time.Second})
 	if err != nil {
 		return errors.Wrapf(err, "failed to open a cache connection to %q", b.dbPath)
 	}
-
+	if b.features.PurgeDb {
-	_ = db.Update(func(tx *bolt.Tx) error {
+		b.Purge()
+	}
+	_ = b.db.Update(func(tx *bolt.Tx) error {
 		_, _ = tx.CreateBucketIfNotExists([]byte(RootBucket))
 		_, _ = tx.CreateBucketIfNotExists([]byte(RootTsBucket))
 		_, _ = tx.CreateBucketIfNotExists([]byte(DataTsBucket))
+		_, _ = tx.CreateBucketIfNotExists([]byte(tempBucket))
 
 		return nil
 	})
 
-	b.db = db
+	b.open = true
 	return nil
 }
 
@@ -136,7 +148,9 @@ func (b *Persistent) getBucket(dir string, createIfMissing bool, tx *bolt.Tx) *b
 	cleanPath(dir)
 
 	entries := strings.FieldsFunc(dir, func(c rune) bool {
-		return os.PathSeparator == c
+		// cover Windows where rclone still uses '/' as path separator
+		// this should be safe as '/' is not a valid Windows character
+		return (os.PathSeparator == c || c == rune('/'))
 	})
 	bucket := tx.Bucket([]byte(RootBucket))
 
@@ -478,6 +492,7 @@ func (b *Persistent) CleanChunksBySize(maxSize int64) {
 	b.cleanupMux.Lock()
 	defer b.cleanupMux.Unlock()
 	var cntChunks int
+	var roughlyCleaned fs.SizeSuffix
 
 	err := b.db.Update(func(tx *bolt.Tx) error {
 		dataTsBucket := tx.Bucket([]byte(DataTsBucket))
@@ -499,6 +514,7 @@ func (b *Persistent) CleanChunksBySize(maxSize int64) {
 
 		if totalSize > maxSize {
 			needToClean := totalSize - maxSize
+			roughlyCleaned = fs.SizeSuffix(needToClean)
 			for k, v := c.First(); k != nil; k, v = c.Next() {
 				var ci chunkInfo
 				err := json.Unmarshal(v, &ci)
@@ -521,7 +537,10 @@ func (b *Persistent) CleanChunksBySize(maxSize int64) {
 				}
 			}
 		}
-		fs.Infof("cache", "deleted (%v) chunks", cntChunks)
+		if cntChunks > 0 {
+			fs.Infof("cache-cleanup", "chunks %v, est. size: %v", cntChunks, roughlyCleaned.String())
+
+		}
 		return nil
 	})
 
@@ -691,6 +710,313 @@ func (b *Persistent) iterateBuckets(buk *bolt.Bucket, bucketFn func(name string)
 	return err
 }
 
+func (b *Persistent) dumpRoot() string {
+	var itBuckets func(buk *bolt.Bucket) map[string]interface{}
+
+	itBuckets = func(buk *bolt.Bucket) map[string]interface{} {
+		m := make(map[string]interface{})
+		c := buk.Cursor()
+		for k, v := c.First(); k != nil; k, v = c.Next() {
+			if v == nil {
+				buk2 := buk.Bucket(k)
+				m[string(k)] = itBuckets(buk2)
+			} else {
+				m[string(k)] = "-"
+			}
+		}
+		return m
+	}
+	var mm map[string]interface{}
+	_ = b.db.View(func(tx *bolt.Tx) error {
+		mm = itBuckets(tx.Bucket([]byte(RootBucket)))
+		return nil
+	})
+	raw, _ := json.MarshalIndent(mm, "", "  ")
+	return string(raw)
+}
+
+// addPendingUpload adds a new file to the pending queue of uploads
+func (b *Persistent) addPendingUpload(destPath string, started bool) error {
+	return b.db.Update(func(tx *bolt.Tx) error {
+		bucket, err := tx.CreateBucketIfNotExists([]byte(tempBucket))
+		if err != nil {
+			return errors.Errorf("couldn't bucket for %v", tempBucket)
+		}
+		tempObj := &tempUploadInfo{
+			DestPath: destPath,
+			AddedOn:  time.Now(),
+			Started:  started,
+		}
+
+		// cache Object Info
+		encoded, err := json.Marshal(tempObj)
+		if err != nil {
+			return errors.Errorf("couldn't marshal object (%v) info: %v", destPath, err)
+		}
+		err = bucket.Put([]byte(destPath), []byte(encoded))
+		if err != nil {
+			return errors.Errorf("couldn't cache object (%v) info: %v", destPath, err)
+		}
+
+		return nil
+	})
+}
+
+// getPendingUpload returns the next file from the pending queue of uploads
+func (b *Persistent) getPendingUpload(inRoot string, waitTime time.Duration) (destPath string, err error) {
+	b.tempQueueMux.Lock()
+	defer b.tempQueueMux.Unlock()
+
+	err = b.db.Update(func(tx *bolt.Tx) error {
+		bucket, err := tx.CreateBucketIfNotExists([]byte(tempBucket))
+		if err != nil {
+			return errors.Errorf("couldn't bucket for %v", tempBucket)
+		}
+
+		c := bucket.Cursor()
+		for k, v := c.Seek([]byte(inRoot)); k != nil && bytes.HasPrefix(k, []byte(inRoot)); k, v = c.Next() {
+			//for k, v := c.First(); k != nil; k, v = c.Next() {
+			var tempObj = &tempUploadInfo{}
+			err = json.Unmarshal(v, tempObj)
+			if err != nil {
+				fs.Errorf(b, "failed to read pending upload: %v", err)
+				continue
+			}
+			// skip over started uploads
+			if tempObj.Started || time.Now().Before(tempObj.AddedOn.Add(waitTime)) {
+				continue
+			}
+
+			tempObj.Started = true
+			v2, err := json.Marshal(tempObj)
+			if err != nil {
+				fs.Errorf(b, "failed to update pending upload: %v", err)
+				continue
+			}
+			err = bucket.Put(k, v2)
+			if err != nil {
+				fs.Errorf(b, "failed to update pending upload: %v", err)
+				continue
+			}
+
+			destPath = tempObj.DestPath
+			return nil
+		}
+
+		return errors.Errorf("no pending upload found")
+	})
+
+	return destPath, err
+}
+
+// SearchPendingUpload returns the file info from the pending queue of uploads
+func (b *Persistent) SearchPendingUpload(remote string) (started bool, err error) {
+	err = b.db.View(func(tx *bolt.Tx) error {
+		bucket := tx.Bucket([]byte(tempBucket))
+		if bucket == nil {
+			return errors.Errorf("couldn't bucket for %v", tempBucket)
+		}
+
+		var tempObj = &tempUploadInfo{}
+		v := bucket.Get([]byte(remote))
+		err = json.Unmarshal(v, tempObj)
+		if err != nil {
+			return errors.Errorf("pending upload (%v) not found %v", remote, err)
+		}
+
+		started = tempObj.Started
+		return nil
+	})
+
+	return started, err
+}
+
+// searchPendingUploadFromDir files currently pending upload from a single dir
+func (b *Persistent) searchPendingUploadFromDir(dir string) (remotes []string, err error) {
+	err = b.db.View(func(tx *bolt.Tx) error {
+		bucket := tx.Bucket([]byte(tempBucket))
+		if bucket == nil {
+			return errors.Errorf("couldn't bucket for %v", tempBucket)
+		}
+
+		c := bucket.Cursor()
+		for k, v := c.First(); k != nil; k, v = c.Next() {
+			var tempObj = &tempUploadInfo{}
+			err = json.Unmarshal(v, tempObj)
+			if err != nil {
+				fs.Errorf(b, "failed to read pending upload: %v", err)
+				continue
+			}
+			parentDir := cleanPath(path.Dir(tempObj.DestPath))
+			if dir == parentDir {
+				remotes = append(remotes, tempObj.DestPath)
+			}
+		}
+
+		return nil
+	})
+
+	return remotes, err
+}
+
+func (b *Persistent) rollbackPendingUpload(remote string) error {
+	b.tempQueueMux.Lock()
+	defer b.tempQueueMux.Unlock()
+
+	return b.db.Update(func(tx *bolt.Tx) error {
+		bucket, err := tx.CreateBucketIfNotExists([]byte(tempBucket))
+		if err != nil {
+			return errors.Errorf("couldn't bucket for %v", tempBucket)
+		}
+		var tempObj = &tempUploadInfo{}
+		v := bucket.Get([]byte(remote))
+		err = json.Unmarshal(v, tempObj)
+		if err != nil {
+			return errors.Errorf("pending upload (%v) not found %v", remote, err)
+		}
+		tempObj.Started = false
+		v2, err := json.Marshal(tempObj)
+		if err != nil {
+			return errors.Errorf("pending upload not updated %v", err)
+		}
+		err = bucket.Put([]byte(tempObj.DestPath), v2)
+		if err != nil {
+			return errors.Errorf("pending upload not updated %v", err)
+		}
+		return nil
+	})
+}
+
+func (b *Persistent) removePendingUpload(remote string) error {
+	b.tempQueueMux.Lock()
+	defer b.tempQueueMux.Unlock()
+
+	return b.db.Update(func(tx *bolt.Tx) error {
+		bucket, err := tx.CreateBucketIfNotExists([]byte(tempBucket))
+		if err != nil {
+			return errors.Errorf("couldn't bucket for %v", tempBucket)
+		}
+		return bucket.Delete([]byte(remote))
+	})
+}
+
+// updatePendingUpload allows to update an existing item in the queue while checking if it's not started in the same
+// transaction. If it is started, it will not allow the update
+func (b *Persistent) updatePendingUpload(remote string, fn func(item *tempUploadInfo) error) error {
+	b.tempQueueMux.Lock()
+	defer b.tempQueueMux.Unlock()
+
+	return b.db.Update(func(tx *bolt.Tx) error {
+		bucket, err := tx.CreateBucketIfNotExists([]byte(tempBucket))
+		if err != nil {
+			return errors.Errorf("couldn't bucket for %v", tempBucket)
+		}
+
+		var tempObj = &tempUploadInfo{}
+		v := bucket.Get([]byte(remote))
+		err = json.Unmarshal(v, tempObj)
+		if err != nil {
+			return errors.Errorf("pending upload (%v) not found %v", remote, err)
+		}
+		if tempObj.Started {
+			return errors.Errorf("pending upload already started %v", remote)
+		}
+		err = fn(tempObj)
+		if err != nil {
+			return err
+		}
+		if remote != tempObj.DestPath {
+			err := bucket.Delete([]byte(remote))
+			if err != nil {
+				return err
+			}
+			// if this is removed then the entry can be removed too
+			if tempObj.DestPath == "" {
+				return nil
+			}
+		}
+		v2, err := json.Marshal(tempObj)
+		if err != nil {
+			return errors.Errorf("pending upload not updated %v", err)
+		}
+		err = bucket.Put([]byte(tempObj.DestPath), v2)
+		if err != nil {
+			return errors.Errorf("pending upload not updated %v", err)
+		}
+
+		return nil
+	})
+}
+
+// SetPendingUploadToStarted is a way to mark an entry as started (even if it's not already)
+// TO BE USED IN TESTING ONLY
+func (b *Persistent) SetPendingUploadToStarted(remote string) error {
+	return b.updatePendingUpload(remote, func(item *tempUploadInfo) error {
+		item.Started = true
+		return nil
+	})
+}
+
+// ReconcileTempUploads will recursively look for all the files in the temp directory and add them to the queue
+func (b *Persistent) ReconcileTempUploads(cacheFs *Fs) error {
+	return b.db.Update(func(tx *bolt.Tx) error {
+		_ = tx.DeleteBucket([]byte(tempBucket))
+		bucket, err := tx.CreateBucketIfNotExists([]byte(tempBucket))
+		if err != nil {
+			return err
+		}
+
+		var queuedEntries []fs.Object
+		err = walk.Walk(cacheFs.tempFs, "", true, -1, func(path string, entries fs.DirEntries, err error) error {
+			for _, o := range entries {
+				if oo, ok := o.(fs.Object); ok {
+					queuedEntries = append(queuedEntries, oo)
+				}
+			}
+			return nil
+		})
+		if err != nil {
+			return err
+		}
+
+		fs.Debugf(cacheFs, "reconciling temporary uploads")
+		for _, queuedEntry := range queuedEntries {
+			destPath := path.Join(cacheFs.Root(), queuedEntry.Remote())
+			tempObj := &tempUploadInfo{
+				DestPath: destPath,
+				AddedOn:  time.Now(),
+				Started:  false,
+			}
+
+			// cache Object Info
+			encoded, err := json.Marshal(tempObj)
+			if err != nil {
+				return errors.Errorf("couldn't marshal object (%v) info: %v", queuedEntry, err)
+			}
+			err = bucket.Put([]byte(destPath), []byte(encoded))
+			if err != nil {
+				return errors.Errorf("couldn't cache object (%v) info: %v", destPath, err)
+			}
+			fs.Debugf(cacheFs, "reconciled temporary upload: %v", destPath)
+		}
+
+		return nil
+	})
+}
+
+// PurgeTempUploads will remove all the pending uploads from the queue
+// TO BE USED IN TESTING ONLY
+func (b *Persistent) PurgeTempUploads() {
+	b.tempQueueMux.Lock()
+	defer b.tempQueueMux.Unlock()
+
+	_ = b.db.Update(func(tx *bolt.Tx) error {
+		_ = tx.DeleteBucket([]byte(tempBucket))
+		_, _ = tx.CreateBucketIfNotExists([]byte(tempBucket))
+		return nil
+	})
+}
+
 // Close should be called when the program ends gracefully
 func (b *Persistent) Close() {
 	b.cleanupMux.Lock()
@@ -700,6 +1026,7 @@ func (b *Persistent) Close() {
 	if err != nil {
 		fs.Errorf(b, "closing handle: %v", err)
 	}
+	b.open = false
 }
 
 // itob returns an 8-byte big endian representation of v.

docs/content/cache.md

@@ -109,12 +109,38 @@ To start a cached mount
 
     rclone mount --allow-other test-cache: /var/tmp/test-cache
 
+### Write Features ###
+
+### Offline uploading ###
+
+In an effort to make writing through cache more reliable, the backend 
+now supports this feature which can be activated by specifying a
+`cache-tmp-upload-path`.
+
+A files goes through these states when using this feature:
+
+1. An upload is started (usually by copying a file on the cache remote)
+2. When the copy to the temporary location is complete the file is part 
+of the cached remote and looks and behaves like any other file (reading included)
+3. After `cache-tmp-wait-time` passes and the file is next in line, `rclone move` 
+is used to move the file to the cloud provider
+4. Reading the file still works during the upload but most modifications on it will be prohibited
+5. Once the move is complete the file is unlocked for modifications as it
+becomes as any other regular file
+6. If the file is being read through `cache` when it's actually
+deleted from the temporary path then `cache` will simply swap the source
+to the cloud provider without interrupting the reading (small blip can happen though)
+
+Files are uploaded in sequence and only one file is uploaded at a time.
+Uploads will be stored in a queue and be processed based on the order they were added.
+The queue and the temporary storage is persistent across restarts and even purges of the cache.
+
 ### Write Support ###
 
 Writes are supported through `cache`.
 One caveat is that a mounted cache remote does not add any retry or fallback
 mechanism to the upload operation. This will depend on the implementation
-of the wrapped remote.
+of the wrapped remote. Consider using `Offline uploading` for reliable writes.
 
 One special case is covered with `cache-writes` which will cache the file
 data at the same time as the upload when it is enabled making it available
@@ -157,6 +183,16 @@ Affected settings:
 
 ### Known issues ###
 
+#### Mount and --dir-cache-time ####
+
+--dir-cache-time controls the first layer of directory caching which works at the mount layer.
+Being an independent caching mechanism from the `cache` backend, it will manage its own entries
+based on the configured time.
+
+To avoid getting in a scenario where dir cache has obsolete data and cache would have the correct
+one, try to set `--dir-cache-time` to a lower time than `--cache-info-age`. Default values are
+already configured in this way. 
+
 #### Windows support - Experimental ####
 
 There are a couple of issues with Windows `mount` functionality that still require some investigations.
@@ -341,3 +377,23 @@ you can enable this flag to have their data stored in the cache store at the
 same time during upload.
 
 **Default**: not set
+
+#### --cache-tmp-upload-path=PATH ####
+
+This is the path where `cache` will use as a temporary storage for new files
+that need to be uploaded to the cloud provider.
+
+Specifying a value will enable this feature. Without it, it is completely disabled
+and files will be uploaded directly to the cloud provider
+
+**Default**: empty
+
+#### --cache-tmp-wait-time=DURATION ####
+
+This is the duration that a file must wait in the temporary location
+_cache-tmp-upload-path_ before it is selected for upload.
+
+Note that only one file is uploaded at a time and it can take longer to
+start the upload if a queue formed for this purpose.
+
+**Default**: 15m