Merge pull request #4864 from MichaelEischer/restore-delete

restore: improve file type mismatch handling
This commit is contained in:
Michael Eischer 2024-06-26 21:27:57 +02:00 committed by GitHub
commit 132965fd5c
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
12 changed files with 427 additions and 93 deletions

View file

@ -20,3 +20,4 @@ https://github.com/restic/restic/issues/407
https://github.com/restic/restic/issues/2662
https://github.com/restic/restic/pull/4837
https://github.com/restic/restic/pull/4838
https://github.com/restic/restic/pull/4864

View file

@ -348,11 +348,6 @@ func (node Node) writeNodeContent(ctx context.Context, repo BlobLoader, f *os.Fi
}
func (node Node) createSymlinkAt(path string) error {
if err := os.Remove(path); err != nil && !errors.Is(err, os.ErrNotExist) {
return errors.Wrap(err, "Symlink")
}
if err := fs.Symlink(node.LinkTarget, path); err != nil {
return errors.WithStack(err)
}

View file

@ -138,6 +138,26 @@ func restoreExtendedAttributes(nodeType, path string, eas []fs.ExtendedAttribute
}
defer closeFileHandle(fileHandle, path) // Replaced inline defer with named function call
// clear old unexpected xattrs by setting them to an empty value
oldEAs, err := fs.GetFileEA(fileHandle)
if err != nil {
return err
}
for _, oldEA := range oldEAs {
found := false
for _, ea := range eas {
if strings.EqualFold(ea.Name, oldEA.Name) {
found = true
break
}
}
if !found {
eas = append(eas, fs.ExtendedAttribute{Name: oldEA.Name, Value: nil})
}
}
if err = fs.SetFileEA(fileHandle, eas); err != nil {
return errors.Errorf("set EA failed for path %v, with: %v", path, err)
}

View file

@ -40,6 +40,11 @@ func setxattr(path, name string, data []byte) error {
return handleXattrErr(xattr.LSet(path, name, data))
}
// removexattr removes the attribute name from path.
func removexattr(path, name string) error {
return handleXattrErr(xattr.LRemove(path, name))
}
func handleXattrErr(err error) error {
switch e := err.(type) {
case nil:
@ -70,12 +75,29 @@ func (node *Node) fillGenericAttributes(_ string, _ os.FileInfo, _ *statT) (allo
}
func (node Node) restoreExtendedAttributes(path string) error {
expectedAttrs := map[string]struct{}{}
for _, attr := range node.ExtendedAttributes {
err := setxattr(path, attr.Name, attr.Value)
if err != nil {
return err
}
expectedAttrs[attr.Name] = struct{}{}
}
// remove unexpected xattrs
xattrs, err := listxattr(path)
if err != nil {
return err
}
for _, name := range xattrs {
if _, ok := expectedAttrs[name]; ok {
continue
}
if err := removexattr(path, name); err != nil {
return err
}
}
return nil
}

View file

@ -0,0 +1,56 @@
//go:build darwin || freebsd || linux || solaris || windows
// +build darwin freebsd linux solaris windows
package restic
import (
"os"
"path/filepath"
"runtime"
"strings"
"testing"
rtest "github.com/restic/restic/internal/test"
)
func setAndVerifyXattr(t *testing.T, file string, attrs []ExtendedAttribute) {
if runtime.GOOS == "windows" {
// windows seems to convert the xattr name to upper case
for i := range attrs {
attrs[i].Name = strings.ToUpper(attrs[i].Name)
}
}
node := Node{
Type: "file",
ExtendedAttributes: attrs,
}
rtest.OK(t, node.restoreExtendedAttributes(file))
nodeActual := Node{
Type: "file",
}
rtest.OK(t, nodeActual.fillExtendedAttributes(file, false))
rtest.Assert(t, nodeActual.sameExtendedAttributes(node), "xattr mismatch got %v expected %v", nodeActual.ExtendedAttributes, node.ExtendedAttributes)
}
func TestOverwriteXattr(t *testing.T) {
dir := t.TempDir()
file := filepath.Join(dir, "file")
rtest.OK(t, os.WriteFile(file, []byte("hello world"), 0o600))
setAndVerifyXattr(t, file, []ExtendedAttribute{
{
Name: "user.foo",
Value: []byte("bar"),
},
})
setAndVerifyXattr(t, file, []ExtendedAttribute{
{
Name: "user.other",
Value: []byte("some"),
},
})
}

View file

@ -1,11 +1,15 @@
package restorer
import (
"fmt"
stdfs "io/fs"
"os"
"sync"
"syscall"
"github.com/cespare/xxhash/v2"
"github.com/restic/restic/internal/debug"
"github.com/restic/restic/internal/errors"
"github.com/restic/restic/internal/fs"
)
@ -39,13 +43,26 @@ func newFilesWriter(count int) *filesWriter {
}
}
func createFile(path string, createSize int64, sparse bool) (*os.File, error) {
f, err := os.OpenFile(path, os.O_CREATE|os.O_WRONLY, 0600)
func openFile(path string) (*os.File, error) {
f, err := fs.OpenFile(path, fs.O_WRONLY|fs.O_NOFOLLOW, 0600)
if err != nil {
if !fs.IsAccessDenied(err) {
return nil, err
}
return nil, err
}
fi, err := f.Stat()
if err != nil {
_ = f.Close()
return nil, err
}
if !fi.Mode().IsRegular() {
_ = f.Close()
return nil, fmt.Errorf("unexpected file type %v at %q", fi.Mode().Type(), path)
}
return f, nil
}
func createFile(path string, createSize int64, sparse bool) (*os.File, error) {
f, err := fs.OpenFile(path, fs.O_CREATE|fs.O_WRONLY|fs.O_NOFOLLOW, 0600)
if err != nil && fs.IsAccessDenied(err) {
// If file is readonly, clear the readonly flag by resetting the
// permissions of the file and try again
// as the metadata will be set again in the second pass and the
@ -53,40 +70,86 @@ func createFile(path string, createSize int64, sparse bool) (*os.File, error) {
if err = fs.ResetPermissions(path); err != nil {
return nil, err
}
if f, err = os.OpenFile(path, os.O_WRONLY, 0600); err != nil {
if f, err = fs.OpenFile(path, fs.O_WRONLY|fs.O_NOFOLLOW, 0600); err != nil {
return nil, err
}
} else if err != nil && (errors.Is(err, syscall.ELOOP) || errors.Is(err, syscall.EISDIR)) {
// symlink or directory, try to remove it later on
f = nil
} else if err != nil {
return nil, err
}
var fi stdfs.FileInfo
if f != nil {
// stat to check that we've opened a regular file
fi, err = f.Stat()
if err != nil {
_ = f.Close()
return nil, err
}
}
if sparse {
err = truncateSparse(f, createSize)
if err != nil {
_ = f.Close()
return nil, err
mustReplace := f == nil || !fi.Mode().IsRegular()
if !mustReplace {
ex := fs.ExtendedStat(fi)
if ex.Links > 1 {
// there is no efficient way to find out which other files might be linked to this file
// thus nuke the existing file and start with a fresh one
mustReplace = true
}
} else {
info, err := f.Stat()
if err != nil {
_ = f.Close()
return nil, err
}
if info.Size() > createSize {
// file is too long must shorten it
err = f.Truncate(createSize)
if err != nil {
_ = f.Close()
}
if mustReplace {
// close handle if we still have it
if f != nil {
if err := f.Close(); err != nil {
return nil, err
}
} else if createSize > 0 {
err := fs.PreallocateFile(f, createSize)
if err != nil {
// Just log the preallocate error but don't let it cause the restore process to fail.
// Preallocate might return an error if the filesystem (implementation) does not
// support preallocation or our parameters combination to the preallocate call
// This should yield a syscall.ENOTSUP error, but some other errors might also
// show up.
debug.Log("Failed to preallocate %v with size %v: %v", path, createSize, err)
}
}
// not what we expected, try to get rid of it
if err := fs.Remove(path); err != nil {
return nil, err
}
// create a new file, pass O_EXCL to make sure there are no surprises
f, err = fs.OpenFile(path, fs.O_CREATE|fs.O_WRONLY|fs.O_EXCL|fs.O_NOFOLLOW, 0600)
if err != nil {
return nil, err
}
fi, err = f.Stat()
if err != nil {
_ = f.Close()
return nil, err
}
}
return ensureSize(f, fi, createSize, sparse)
}
func ensureSize(f *os.File, fi stdfs.FileInfo, createSize int64, sparse bool) (*os.File, error) {
if sparse {
err := truncateSparse(f, createSize)
if err != nil {
_ = f.Close()
return nil, err
}
} else if fi.Size() > createSize {
// file is too long must shorten it
err := f.Truncate(createSize)
if err != nil {
_ = f.Close()
return nil, err
}
} else if createSize > 0 {
err := fs.PreallocateFile(f, createSize)
if err != nil {
// Just log the preallocate error but don't let it cause the restore process to fail.
// Preallocate might return an error if the filesystem (implementation) does not
// support preallocation or our parameters combination to the preallocate call
// This should yield a syscall.ENOTSUP error, but some other errors might also
// show up.
debug.Log("Failed to preallocate %v with size %v: %v", f.Name(), createSize, err)
}
}
return f, nil
@ -110,7 +173,7 @@ func (w *filesWriter) writeToFile(path string, blob []byte, offset int64, create
if err != nil {
return nil, err
}
} else if f, err = os.OpenFile(path, os.O_WRONLY, 0600); err != nil {
} else if f, err = openFile(path); err != nil {
return nil, err
}

View file

@ -0,0 +1,10 @@
//go:build !windows
// +build !windows
package restorer
import "syscall"
func notEmptyDirError() error {
return syscall.ENOTEMPTY
}

View file

@ -1,9 +1,13 @@
package restorer
import (
"fmt"
"os"
"path/filepath"
"runtime"
"testing"
"github.com/restic/restic/internal/errors"
rtest "github.com/restic/restic/internal/test"
)
@ -34,3 +38,94 @@ func TestFilesWriterBasic(t *testing.T) {
rtest.OK(t, err)
rtest.Equals(t, []byte{2, 2}, buf)
}
func TestCreateFile(t *testing.T) {
basepath := filepath.Join(t.TempDir(), "test")
scenarios := []struct {
name string
create func(t testing.TB, path string)
check func(t testing.TB, path string)
err error
}{
{
name: "file",
create: func(t testing.TB, path string) {
rtest.OK(t, os.WriteFile(path, []byte("test-test-test-data"), 0o400))
},
},
{
name: "empty dir",
create: func(t testing.TB, path string) {
rtest.OK(t, os.Mkdir(path, 0o400))
},
},
{
name: "symlink",
create: func(t testing.TB, path string) {
rtest.OK(t, os.Symlink("./something", path))
},
},
{
name: "filled dir",
create: func(t testing.TB, path string) {
rtest.OK(t, os.Mkdir(path, 0o700))
rtest.OK(t, os.WriteFile(filepath.Join(path, "file"), []byte("data"), 0o400))
},
err: notEmptyDirError(),
},
{
name: "hardlinks",
create: func(t testing.TB, path string) {
rtest.OK(t, os.WriteFile(path, []byte("test-test-test-data"), 0o400))
rtest.OK(t, os.Link(path, path+"h"))
},
check: func(t testing.TB, path string) {
if runtime.GOOS == "windows" {
// hardlinks are not supported on windows
return
}
data, err := os.ReadFile(path + "h")
rtest.OK(t, err)
rtest.Equals(t, "test-test-test-data", string(data), "unexpected content change")
},
},
}
tests := []struct {
size int64
isSparse bool
}{
{5, false},
{21, false},
{100, false},
{5, true},
{21, true},
{100, true},
}
for i, sc := range scenarios {
t.Run(sc.name, func(t *testing.T) {
for j, test := range tests {
path := basepath + fmt.Sprintf("%v%v", i, j)
sc.create(t, path)
f, err := createFile(path, test.size, test.isSparse)
if sc.err == nil {
rtest.OK(t, err)
fi, err := f.Stat()
rtest.OK(t, err)
rtest.Assert(t, fi.Mode().IsRegular(), "wrong filetype %v", fi.Mode())
rtest.Assert(t, fi.Size() <= test.size, "unexpected file size expected %v, got %v", test.size, fi.Size())
rtest.OK(t, f.Close())
if sc.check != nil {
sc.check(t, path)
}
} else {
rtest.Assert(t, errors.Is(err, sc.err), "unexpected error got %v expected %v", err, sc.err)
}
rtest.OK(t, os.RemoveAll(path))
}
})
}
}

View file

@ -0,0 +1,7 @@
package restorer
import "syscall"
func notEmptyDirError() error {
return syscall.ERROR_DIR_NOT_EMPTY
}

View file

@ -221,6 +221,9 @@ func (res *Restorer) traverseTree(ctx context.Context, target, location string,
func (res *Restorer) restoreNodeTo(ctx context.Context, node *restic.Node, target, location string) error {
debug.Log("restoreNode %v %v %v", node.Name, target, location)
if err := fs.Remove(target); err != nil && !errors.Is(err, os.ErrNotExist) {
return errors.Wrap(err, "RemoveNode")
}
err := node.CreateAt(ctx, target, res.repo)
if err != nil {
@ -242,7 +245,7 @@ func (res *Restorer) restoreNodeMetadataTo(node *restic.Node, target, location s
}
func (res *Restorer) restoreHardlinkAt(node *restic.Node, target, path, location string) error {
if err := fs.Remove(path); !os.IsNotExist(err) {
if err := fs.Remove(path); err != nil && !errors.Is(err, os.ErrNotExist) {
return errors.Wrap(err, "RemoveCreateHardlink")
}
err := fs.Link(target, path)
@ -256,6 +259,23 @@ func (res *Restorer) restoreHardlinkAt(node *restic.Node, target, path, location
return res.restoreNodeMetadataTo(node, path, location)
}
func (res *Restorer) ensureDir(target string) error {
fi, err := fs.Lstat(target)
if err != nil && !errors.Is(err, os.ErrNotExist) {
return fmt.Errorf("failed to check for directory: %w", err)
}
if err == nil && !fi.IsDir() {
// try to cleanup unexpected file
if err := fs.Remove(target); err != nil {
return fmt.Errorf("failed to remove stale item: %w", err)
}
}
// create parent dir with default permissions
// second pass #leaveDir restores dir metadata after visiting/restoring all children
return fs.MkdirAll(target, 0700)
}
// RestoreTo creates the directories and files in the snapshot below dst.
// Before an item is created, res.Filter is called.
func (res *Restorer) RestoreTo(ctx context.Context, dst string) error {
@ -281,17 +301,12 @@ func (res *Restorer) RestoreTo(ctx context.Context, dst string) error {
enterDir: func(_ *restic.Node, target, location string) error {
debug.Log("first pass, enterDir: mkdir %q, leaveDir should restore metadata", location)
res.opts.Progress.AddFile(0)
// create dir with default permissions
// #leaveDir restores dir metadata after visiting all children
return fs.MkdirAll(target, 0700)
return res.ensureDir(target)
},
visitNode: func(node *restic.Node, target, location string) error {
debug.Log("first pass, visitNode: mkdir %q, leaveDir on second pass should restore metadata", location)
// create parent dir with default permissions
// second pass #leaveDir restores dir metadata after visiting/restoring all children
err := fs.MkdirAll(filepath.Dir(target), 0700)
if err != nil {
if err := res.ensureDir(filepath.Dir(target)); err != nil {
return err
}
@ -526,7 +541,7 @@ func (s *fileState) HasMatchingBlob(i int) bool {
// Reusing buffers prevents the verifier goroutines allocating all of RAM and
// flushing the filesystem cache (at least on Linux).
func (res *Restorer) verifyFile(target string, node *restic.Node, failFast bool, trustMtime bool, buf []byte) (*fileState, []byte, error) {
f, err := os.OpenFile(target, fs.O_RDONLY|fs.O_NOFOLLOW, 0)
f, err := fs.OpenFile(target, fs.O_RDONLY|fs.O_NOFOLLOW, 0)
if err != nil {
return nil, buf, err
}

View file

@ -37,6 +37,11 @@ type File struct {
attributes *FileAttributes
}
type Symlink struct {
Target string
ModTime time.Time
}
type Dir struct {
Nodes map[string]Node
Mode os.FileMode
@ -103,6 +108,20 @@ func saveDir(t testing.TB, repo restic.BlobSaver, nodes map[string]Node, inode u
GenericAttributes: getGenericAttributes(node.attributes, false),
})
rtest.OK(t, err)
case Symlink:
symlink := n.(Symlink)
err := tree.Insert(&restic.Node{
Type: "symlink",
Mode: os.ModeSymlink | 0o777,
ModTime: symlink.ModTime,
Name: name,
UID: uint32(os.Getuid()),
GID: uint32(os.Getgid()),
LinkTarget: symlink.Target,
Inode: inode,
Links: 1,
})
rtest.OK(t, err)
case Dir:
id := saveDir(t, repo, node.Nodes, inode, getGenericAttributes)
@ -895,6 +914,31 @@ func TestRestorerSparseFiles(t *testing.T) {
len(zeros), blocks, 100*sparsity)
}
func saveSnapshotsAndOverwrite(t *testing.T, baseSnapshot Snapshot, overwriteSnapshot Snapshot, options Options) string {
repo := repository.TestRepository(t)
tempdir := filepath.Join(rtest.TempDir(t), "target")
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
// base snapshot
sn, id := saveSnapshot(t, repo, baseSnapshot, noopGetGenericAttributes)
t.Logf("base snapshot saved as %v", id.Str())
res := NewRestorer(repo, sn, options)
rtest.OK(t, res.RestoreTo(ctx, tempdir))
// overwrite snapshot
sn, id = saveSnapshot(t, repo, overwriteSnapshot, noopGetGenericAttributes)
t.Logf("overwrite snapshot saved as %v", id.Str())
res = NewRestorer(repo, sn, options)
rtest.OK(t, res.RestoreTo(ctx, tempdir))
_, err := res.VerifyFiles(ctx, tempdir)
rtest.OK(t, err)
return tempdir
}
func TestRestorerSparseOverwrite(t *testing.T) {
baseSnapshot := Snapshot{
Nodes: map[string]Node{
@ -908,29 +952,7 @@ func TestRestorerSparseOverwrite(t *testing.T) {
},
}
repo := repository.TestRepository(t)
tempdir := filepath.Join(rtest.TempDir(t), "target")
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
// base snapshot
sn, id := saveSnapshot(t, repo, baseSnapshot, noopGetGenericAttributes)
t.Logf("base snapshot saved as %v", id.Str())
res := NewRestorer(repo, sn, Options{Sparse: true})
err := res.RestoreTo(ctx, tempdir)
rtest.OK(t, err)
// sparse snapshot
sn, id = saveSnapshot(t, repo, sparseSnapshot, noopGetGenericAttributes)
t.Logf("base snapshot saved as %v", id.Str())
res = NewRestorer(repo, sn, Options{Sparse: true, Overwrite: OverwriteAlways})
err = res.RestoreTo(ctx, tempdir)
rtest.OK(t, err)
files, err := res.VerifyFiles(ctx, tempdir)
rtest.OK(t, err)
rtest.Equals(t, 1, files, "unexpected number of verified files")
saveSnapshotsAndOverwrite(t, baseSnapshot, sparseSnapshot, Options{Sparse: true, Overwrite: OverwriteAlways})
}
func TestRestorerOverwriteBehavior(t *testing.T) {
@ -993,26 +1015,7 @@ func TestRestorerOverwriteBehavior(t *testing.T) {
for _, test := range tests {
t.Run("", func(t *testing.T) {
repo := repository.TestRepository(t)
tempdir := filepath.Join(rtest.TempDir(t), "target")
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
// base snapshot
sn, id := saveSnapshot(t, repo, baseSnapshot, noopGetGenericAttributes)
t.Logf("base snapshot saved as %v", id.Str())
res := NewRestorer(repo, sn, Options{})
rtest.OK(t, res.RestoreTo(ctx, tempdir))
// overwrite snapshot
sn, id = saveSnapshot(t, repo, overwriteSnapshot, noopGetGenericAttributes)
t.Logf("overwrite snapshot saved as %v", id.Str())
res = NewRestorer(repo, sn, Options{Overwrite: test.Overwrite})
rtest.OK(t, res.RestoreTo(ctx, tempdir))
_, err := res.VerifyFiles(ctx, tempdir)
rtest.OK(t, err)
tempdir := saveSnapshotsAndOverwrite(t, baseSnapshot, overwriteSnapshot, Options{Overwrite: test.Overwrite})
for filename, content := range test.Files {
data, err := os.ReadFile(filepath.Join(tempdir, filepath.FromSlash(filename)))
@ -1029,6 +1032,56 @@ func TestRestorerOverwriteBehavior(t *testing.T) {
}
}
func TestRestorerOverwriteSpecial(t *testing.T) {
baseTime := time.Now()
baseSnapshot := Snapshot{
Nodes: map[string]Node{
"dirtest": Dir{ModTime: baseTime},
"link": Symlink{Target: "foo", ModTime: baseTime},
"file": File{Data: "content: file\n", Inode: 42, Links: 2, ModTime: baseTime},
"hardlink": File{Data: "content: file\n", Inode: 42, Links: 2, ModTime: baseTime},
"newdir": File{Data: "content: dir\n", ModTime: baseTime},
},
}
overwriteSnapshot := Snapshot{
Nodes: map[string]Node{
"dirtest": Symlink{Target: "foo", ModTime: baseTime},
"link": File{Data: "content: link\n", Inode: 42, Links: 2, ModTime: baseTime.Add(time.Second)},
"file": Symlink{Target: "foo2", ModTime: baseTime},
"hardlink": File{Data: "content: link\n", Inode: 42, Links: 2, ModTime: baseTime.Add(time.Second)},
"newdir": Dir{ModTime: baseTime},
},
}
files := map[string]string{
"link": "content: link\n",
"hardlink": "content: link\n",
}
links := map[string]string{
"dirtest": "foo",
"file": "foo2",
}
tempdir := saveSnapshotsAndOverwrite(t, baseSnapshot, overwriteSnapshot, Options{Overwrite: OverwriteAlways})
for filename, content := range files {
data, err := os.ReadFile(filepath.Join(tempdir, filepath.FromSlash(filename)))
if err != nil {
t.Errorf("unable to read file %v: %v", filename, err)
continue
}
if !bytes.Equal(data, []byte(content)) {
t.Errorf("file %v has wrong content: want %q, got %q", filename, content, data)
}
}
for filename, target := range links {
link, err := fs.Readlink(filepath.Join(tempdir, filepath.FromSlash(filename)))
rtest.OK(t, err)
rtest.Equals(t, link, target, "wrong symlink target")
}
}
func TestRestoreModified(t *testing.T) {
// overwrite files between snapshots and also change their filesize
snapshots := []Snapshot{

View file

@ -1,6 +1,3 @@
//go:build !windows
// +build !windows
package restorer
import (