//
// SPDX-License-Identifier: AGPL-3.0
-package main
+package keepstore
import (
- "bufio"
"context"
- "flag"
+ "encoding/json"
+ "errors"
"fmt"
"io"
"io/ioutil"
"syscall"
"time"
+ "git.arvados.org/arvados.git/sdk/go/arvados"
"github.com/prometheus/client_golang/prometheus"
+ "github.com/sirupsen/logrus"
)
-type unixVolumeAdder struct {
- *Config
-}
-
-// String implements flag.Value
-func (vs *unixVolumeAdder) String() string {
- return "-"
+func init() {
+ driver["Directory"] = newDirectoryVolume
}
-func (vs *unixVolumeAdder) Set(path string) error {
- if dirs := strings.Split(path, ","); len(dirs) > 1 {
- log.Print("DEPRECATED: using comma-separated volume list.")
- for _, dir := range dirs {
- if err := vs.Set(dir); err != nil {
- return err
- }
- }
- return nil
+func newDirectoryVolume(cluster *arvados.Cluster, volume arvados.Volume, logger logrus.FieldLogger, metrics *volumeMetricsVecs) (Volume, error) {
+ v := &UnixVolume{cluster: cluster, volume: volume, logger: logger, metrics: metrics}
+ err := json.Unmarshal(volume.DriverParameters, &v)
+ if err != nil {
+ return nil, err
}
- vs.Config.Volumes = append(vs.Config.Volumes, &UnixVolume{
- Root: path,
- ReadOnly: deprecated.flagReadonly,
- Serialize: deprecated.flagSerializeIO,
- })
- return nil
+ v.logger = v.logger.WithField("Volume", v.String())
+ return v, v.check()
}
-func init() {
- VolumeTypes = append(VolumeTypes, func() VolumeWithExamples { return &UnixVolume{} })
-
- flag.Var(&unixVolumeAdder{theConfig}, "volumes", "see Volumes configuration")
- flag.Var(&unixVolumeAdder{theConfig}, "volume", "see Volumes configuration")
-}
-
-// Discover adds a UnixVolume for every directory named "keep" that is
-// located at the top level of a device- or tmpfs-backed mount point
-// other than "/". It returns the number of volumes added.
-func (vs *unixVolumeAdder) Discover() int {
- added := 0
- f, err := os.Open(ProcMounts)
- if err != nil {
- log.Fatalf("opening %s: %s", ProcMounts, err)
+func (v *UnixVolume) check() error {
+ if v.Root == "" {
+ return errors.New("DriverParameters.Root was not provided")
}
- scanner := bufio.NewScanner(f)
- for scanner.Scan() {
- args := strings.Fields(scanner.Text())
- if err := scanner.Err(); err != nil {
- log.Fatalf("reading %s: %s", ProcMounts, err)
- }
- dev, mount := args[0], args[1]
- if mount == "/" {
- continue
- }
- if dev != "tmpfs" && !strings.HasPrefix(dev, "/dev/") {
- continue
- }
- keepdir := mount + "/keep"
- if st, err := os.Stat(keepdir); err != nil || !st.IsDir() {
- continue
- }
- // Set the -readonly flag (but only for this volume)
- // if the filesystem is mounted readonly.
- flagReadonlyWas := deprecated.flagReadonly
- for _, fsopt := range strings.Split(args[3], ",") {
- if fsopt == "ro" {
- deprecated.flagReadonly = true
- break
- }
- if fsopt == "rw" {
- break
- }
- }
- if err := vs.Set(keepdir); err != nil {
- log.Printf("adding %q: %s", keepdir, err)
- } else {
- added++
- }
- deprecated.flagReadonly = flagReadonlyWas
+ if v.Serialize {
+ v.locker = &sync.Mutex{}
}
- return added
+ if !strings.HasPrefix(v.Root, "/") {
+ return fmt.Errorf("DriverParameters.Root %q does not start with '/'", v.Root)
+ }
+
+ // Set up prometheus metrics
+ lbls := prometheus.Labels{"device_id": v.GetDeviceID()}
+ v.os.stats.opsCounters, v.os.stats.errCounters, v.os.stats.ioBytes = v.metrics.getCounterVecsFor(lbls)
+
+ _, err := v.os.Stat(v.Root)
+ return err
}
// A UnixVolume stores and retrieves blocks in a local directory.
type UnixVolume struct {
- Root string // path to the volume's root directory
- ReadOnly bool
- Serialize bool
- DirectoryReplication int
- StorageClasses []string
+ Root string // path to the volume's root directory
+ Serialize bool
+
+ cluster *arvados.Cluster
+ volume arvados.Volume
+ logger logrus.FieldLogger
+ metrics *volumeMetricsVecs
// something to lock during IO, typically a sync.Mutex (or nil
// to skip locking)
os osWithStats
}
-// DeviceID returns a globally unique ID for the volume's root
+// GetDeviceID returns a globally unique ID for the volume's root
// directory, consisting of the filesystem's UUID and the path from
// filesystem root to storage directory, joined by "/". For example,
-// the DeviceID for a local directory "/mnt/xvda1/keep" might be
+// the device ID for a local directory "/mnt/xvda1/keep" might be
// "fa0b6166-3b55-4994-bd3f-92f4e00a1bb0/keep".
-func (v *UnixVolume) DeviceID() string {
+func (v *UnixVolume) GetDeviceID() string {
giveup := func(f string, args ...interface{}) string {
- log.Printf(f+"; using blank DeviceID for volume %s", append(args, v)...)
+ v.logger.Infof(f+"; using blank DeviceID for volume %s", append(args, v)...)
return ""
}
buf, err := exec.Command("findmnt", "--noheadings", "--target", v.Root).CombinedOutput()
fi, err := os.Stat(dev)
if err != nil {
- return giveup("stat %q: %s\n", dev, err)
+ return giveup("stat %q: %s", dev, err)
}
ino := fi.Sys().(*syscall.Stat_t).Ino
if err != nil {
return giveup("opening %q: %s", udir, err)
}
+ defer d.Close()
uuids, err := d.Readdirnames(0)
if err != nil {
return giveup("reading %q: %s", udir, err)
link := filepath.Join(udir, uuid)
fi, err = os.Stat(link)
if err != nil {
- log.Printf("error: stat %q: %s", link, err)
+ v.logger.WithError(err).Errorf("stat(%q) failed", link)
continue
}
if fi.Sys().(*syscall.Stat_t).Ino == ino {
return giveup("could not find entry in %q matching %q", udir, dev)
}
-// Examples implements VolumeWithExamples.
-func (*UnixVolume) Examples() []Volume {
- return []Volume{
- &UnixVolume{
- Root: "/mnt/local-disk",
- Serialize: true,
- DirectoryReplication: 1,
- },
- &UnixVolume{
- Root: "/mnt/network-disk",
- Serialize: false,
- DirectoryReplication: 2,
- },
- }
-}
-
-// Type implements Volume
-func (v *UnixVolume) Type() string {
- return "Directory"
-}
-
-// Start implements Volume
-func (v *UnixVolume) Start(vm *volumeMetricsVecs) error {
- if v.Serialize {
- v.locker = &sync.Mutex{}
- }
- if !strings.HasPrefix(v.Root, "/") {
- return fmt.Errorf("volume root does not start with '/': %q", v.Root)
- }
- if v.DirectoryReplication == 0 {
- v.DirectoryReplication = 1
- }
- // Set up prometheus metrics
- lbls := prometheus.Labels{"device_id": v.DeviceID()}
- v.os.stats.opsCounters, v.os.stats.errCounters, v.os.stats.ioBytes = vm.getCounterVecsFor(lbls)
-
- _, err := v.os.Stat(v.Root)
-
- return err
-}
-
// Touch sets the timestamp for the given locator to the current time
func (v *UnixVolume) Touch(loc string) error {
- if v.ReadOnly {
+ if v.volume.ReadOnly {
return MethodDisabledError
}
p := v.blockPath(loc)
return e
}
defer v.unlockfile(f)
- ts := syscall.NsecToTimespec(time.Now().UnixNano())
+ ts := time.Now()
v.os.stats.TickOps("utimes")
v.os.stats.Tick(&v.os.stats.UtimesOps)
- err = syscall.UtimesNano(p, []syscall.Timespec{ts, ts})
+ err = os.Chtimes(p, ts, ts)
v.os.stats.TickErr(err)
return err
}
// WriteBlock implements BlockWriter.
func (v *UnixVolume) WriteBlock(ctx context.Context, loc string, rdr io.Reader) error {
- if v.ReadOnly {
+ if v.volume.ReadOnly {
return MethodDisabledError
}
if v.IsFull() {
}
bdir := v.blockDir(loc)
if err := os.MkdirAll(bdir, 0755); err != nil {
- log.Printf("%s: could not create directory %s: %s",
- loc, bdir, err)
- return err
- }
-
- tmpfile, tmperr := v.os.TempFile(bdir, "tmp"+loc)
- if tmperr != nil {
- log.Printf("ioutil.TempFile(%s, tmp%s): %s", bdir, loc, tmperr)
- return tmperr
+ return fmt.Errorf("error creating directory %s: %s", bdir, err)
}
bpath := v.blockPath(loc)
+ tmpfile, err := v.os.TempFile(bdir, "tmp"+loc)
+ if err != nil {
+ return fmt.Errorf("TempFile(%s, tmp%s) failed: %s", bdir, loc, err)
+ }
+ defer v.os.Remove(tmpfile.Name())
+ defer tmpfile.Close()
- if err := v.lock(ctx); err != nil {
+ if err = v.lock(ctx); err != nil {
return err
}
defer v.unlock()
n, err := io.Copy(tmpfile, rdr)
v.os.stats.TickOutBytes(uint64(n))
if err != nil {
- log.Printf("%s: writing to %s: %s\n", v, bpath, err)
- tmpfile.Close()
- v.os.Remove(tmpfile.Name())
- return err
+ return fmt.Errorf("error writing %s: %s", bpath, err)
}
- if err := tmpfile.Close(); err != nil {
- log.Printf("closing %s: %s\n", tmpfile.Name(), err)
- v.os.Remove(tmpfile.Name())
- return err
+ if err = tmpfile.Close(); err != nil {
+ return fmt.Errorf("error closing %s: %s", tmpfile.Name(), err)
+ }
+ // ext4 uses a low-precision clock and effectively backdates
+ // files by up to 10 ms, sometimes across a 1-second boundary,
+ // which produces confusing results in logs and tests. We
+ // avoid this by setting the output file's timestamps
+ // explicitly, using a higher resolution clock.
+ ts := time.Now()
+ v.os.stats.TickOps("utimes")
+ v.os.stats.Tick(&v.os.stats.UtimesOps)
+ if err = os.Chtimes(tmpfile.Name(), ts, ts); err != nil {
+ return fmt.Errorf("error setting timestamps on %s: %s", tmpfile.Name(), err)
}
- if err := v.os.Rename(tmpfile.Name(), bpath); err != nil {
- log.Printf("rename %s %s: %s\n", tmpfile.Name(), bpath, err)
- return v.os.Remove(tmpfile.Name())
+ if err = v.os.Rename(tmpfile.Name(), bpath); err != nil {
+ return fmt.Errorf("error renaming %s to %s: %s", tmpfile.Name(), bpath, err)
}
return nil
}
func (v *UnixVolume) Status() *VolumeStatus {
fi, err := v.os.Stat(v.Root)
if err != nil {
- log.Printf("%s: os.Stat: %s\n", v, err)
+ v.logger.WithError(err).Error("stat failed")
return nil
}
- devnum := fi.Sys().(*syscall.Stat_t).Dev
+ // uint64() cast here supports GOOS=darwin where Dev is
+ // int32. If the device number is negative, the unsigned
+ // devnum won't be the real device number any more, but that's
+ // fine -- all we care about is getting the same number each
+ // time.
+ devnum := uint64(fi.Sys().(*syscall.Stat_t).Dev)
var fs syscall.Statfs_t
if err := syscall.Statfs(v.Root, &fs); err != nil {
- log.Printf("%s: statfs: %s\n", v, err)
+ v.logger.WithError(err).Error("statfs failed")
return nil
}
// These calculations match the way df calculates disk usage:
// e4de7a2810f5554cd39b36d8ddb132ff+67108864 1388701136
//
func (v *UnixVolume) IndexTo(prefix string, w io.Writer) error {
- var lastErr error
rootdir, err := v.os.Open(v.Root)
if err != nil {
return err
}
- defer rootdir.Close()
v.os.stats.TickOps("readdir")
v.os.stats.Tick(&v.os.stats.ReaddirOps)
- for {
- names, err := rootdir.Readdirnames(1)
- if err == io.EOF {
- return lastErr
- } else if err != nil {
- return err
- }
- if !strings.HasPrefix(names[0], prefix) && !strings.HasPrefix(prefix, names[0]) {
+ subdirs, err := rootdir.Readdirnames(-1)
+ rootdir.Close()
+ if err != nil {
+ return err
+ }
+ for _, subdir := range subdirs {
+ if !strings.HasPrefix(subdir, prefix) && !strings.HasPrefix(prefix, subdir) {
// prefix excludes all blocks stored in this dir
continue
}
- if !blockDirRe.MatchString(names[0]) {
- continue
- }
- blockdirpath := filepath.Join(v.Root, names[0])
- blockdir, err := v.os.Open(blockdirpath)
- if err != nil {
- log.Print("Error reading ", blockdirpath, ": ", err)
- lastErr = err
+ if !blockDirRe.MatchString(subdir) {
continue
}
- v.os.stats.TickOps("readdir")
- v.os.stats.Tick(&v.os.stats.ReaddirOps)
- for {
- fileInfo, err := blockdir.Readdir(1)
- if err == io.EOF {
+ blockdirpath := filepath.Join(v.Root, subdir)
+
+ var dirents []os.DirEntry
+ for attempt := 0; ; attempt++ {
+ v.os.stats.TickOps("readdir")
+ v.os.stats.Tick(&v.os.stats.ReaddirOps)
+ dirents, err = os.ReadDir(blockdirpath)
+ if err == nil {
break
+ } else if attempt < 5 && strings.Contains(err.Error(), "errno 523") {
+ // EBADCOOKIE (NFS stopped accepting
+ // our readdirent cookie) -- retry a
+ // few times before giving up
+ v.logger.WithError(err).Printf("retry after error reading %s", blockdirpath)
+ continue
+ } else {
+ return err
+ }
+ }
+
+ for _, dirent := range dirents {
+ fileInfo, err := dirent.Info()
+ if os.IsNotExist(err) {
+ // File disappeared between ReadDir() and now
+ continue
} else if err != nil {
- log.Print("Error reading ", blockdirpath, ": ", err)
- lastErr = err
- break
+ v.logger.WithError(err).Errorf("error getting FileInfo for %q in %q", dirent.Name(), blockdirpath)
+ return err
}
- name := fileInfo[0].Name()
+ name := fileInfo.Name()
if !strings.HasPrefix(name, prefix) {
continue
}
}
_, err = fmt.Fprint(w,
name,
- "+", fileInfo[0].Size(),
- " ", fileInfo[0].ModTime().UnixNano(),
+ "+", fileInfo.Size(),
+ " ", fileInfo.ModTime().UnixNano(),
"\n")
if err != nil {
- log.Print("Error writing : ", err)
- lastErr = err
- break
+ return fmt.Errorf("error writing: %s", err)
}
}
- blockdir.Close()
}
+ return nil
}
// Trash trashes the block data from the unix storage
-// If TrashLifetime == 0, the block is deleted
+// If BlobTrashLifetime == 0, the block is deleted
// Else, the block is renamed as path/{loc}.trash.{deadline},
-// where deadline = now + TrashLifetime
+// where deadline = now + BlobTrashLifetime
func (v *UnixVolume) Trash(loc string) error {
// Touch() must be called before calling Write() on a block. Touch()
// also uses lockfile(). This avoids a race condition between Write()
// Trash() will read the correct up-to-date timestamp and choose not to
// trash the file.
- if v.ReadOnly {
+ if v.volume.ReadOnly || !v.cluster.Collections.BlobTrash {
return MethodDisabledError
}
if err := v.lock(context.TODO()); err != nil {
// anyway (because the permission signatures have expired).
if fi, err := v.os.Stat(p); err != nil {
return err
- } else if time.Since(fi.ModTime()) < time.Duration(theConfig.BlobSignatureTTL) {
+ } else if time.Since(fi.ModTime()) < v.cluster.Collections.BlobSigningTTL.Duration() {
return nil
}
- if theConfig.TrashLifetime == 0 {
+ if v.cluster.Collections.BlobTrashLifetime == 0 {
return v.os.Remove(p)
}
- return v.os.Rename(p, fmt.Sprintf("%v.trash.%d", p, time.Now().Add(theConfig.TrashLifetime.Duration()).Unix()))
+ return v.os.Rename(p, fmt.Sprintf("%v.trash.%d", p, time.Now().Add(v.cluster.Collections.BlobTrashLifetime.Duration()).Unix()))
}
// Untrash moves block from trash back into store
// Look for path/{loc}.trash.{deadline} in storage,
// and rename the first such file as path/{loc}
func (v *UnixVolume) Untrash(loc string) (err error) {
- if v.ReadOnly {
+ if v.volume.ReadOnly {
return MethodDisabledError
}
if avail, err := v.FreeDiskSpace(); err == nil {
isFull = avail < MinFreeKilobytes
} else {
- log.Printf("%s: FreeDiskSpace: %s\n", v, err)
+ v.logger.WithError(err).Errorf("%s: FreeDiskSpace failed", v)
isFull = false
}
return fmt.Sprintf("[UnixVolume %s]", v.Root)
}
-// Writable returns false if all future Put, Mtime, and Delete calls
-// are expected to fail.
-func (v *UnixVolume) Writable() bool {
- return !v.ReadOnly
-}
-
-// Replication returns the number of replicas promised by the
-// underlying device (as specified in configuration).
-func (v *UnixVolume) Replication() int {
- return v.DirectoryReplication
-}
-
-// GetStorageClasses implements Volume
-func (v *UnixVolume) GetStorageClasses() []string {
- return v.StorageClasses
-}
-
// InternalStats returns I/O and filesystem ops counters.
func (v *UnixVolume) InternalStats() interface{} {
return &v.os.stats
if v.locker == nil {
return nil
}
+ t0 := time.Now()
locked := make(chan struct{})
go func() {
v.locker.Lock()
}()
select {
case <-ctx.Done():
+ v.logger.Infof("client hung up while waiting for Serialize lock (%s)", time.Since(t0))
go func() {
<-locked
v.locker.Unlock()
// EmptyTrash walks hierarchy looking for {hash}.trash.*
// and deletes those with deadline < now.
func (v *UnixVolume) EmptyTrash() {
+ if v.cluster.Collections.BlobDeleteConcurrency < 1 {
+ return
+ }
+
var bytesDeleted, bytesInTrash int64
var blocksDeleted, blocksInTrash int64
}
deadline, err := strconv.ParseInt(matches[2], 10, 64)
if err != nil {
- log.Printf("EmptyTrash: %v: ParseInt(%v): %v", path, matches[2], err)
+ v.logger.WithError(err).Errorf("EmptyTrash: %v: ParseInt(%q) failed", path, matches[2])
return
}
atomic.AddInt64(&bytesInTrash, info.Size())
}
err = v.os.Remove(path)
if err != nil {
- log.Printf("EmptyTrash: Remove %v: %v", path, err)
+ v.logger.WithError(err).Errorf("EmptyTrash: Remove(%q) failed", path)
return
}
atomic.AddInt64(&bytesDeleted, info.Size())
info os.FileInfo
}
var wg sync.WaitGroup
- todo := make(chan dirent, theConfig.EmptyTrashWorkers)
- for i := 0; i < 1 || i < theConfig.EmptyTrashWorkers; i++ {
+ todo := make(chan dirent, v.cluster.Collections.BlobDeleteConcurrency)
+ for i := 0; i < v.cluster.Collections.BlobDeleteConcurrency; i++ {
wg.Add(1)
go func() {
defer wg.Done()
err := filepath.Walk(v.Root, func(path string, info os.FileInfo, err error) error {
if err != nil {
- log.Printf("EmptyTrash: filepath.Walk: %v: %v", path, err)
+ v.logger.WithError(err).Errorf("EmptyTrash: filepath.Walk(%q) failed", path)
+ // Don't give up -- keep walking other
+ // files/dirs
+ return nil
+ } else if !info.Mode().IsDir() {
+ todo <- dirent{path, info}
return nil
+ } else if path == v.Root || blockDirRe.MatchString(info.Name()) {
+ // Descend into a directory that we might have
+ // put trash in.
+ return nil
+ } else {
+ // Don't descend into other dirs.
+ return filepath.SkipDir
}
- todo <- dirent{path, info}
- return nil
})
close(todo)
wg.Wait()
if err != nil {
- log.Printf("EmptyTrash error for %v: %v", v.String(), err)
+ v.logger.WithError(err).Error("EmptyTrash failed")
}
- log.Printf("EmptyTrash stats for %v: Deleted %v bytes in %v blocks. Remaining in trash: %v bytes in %v blocks.", v.String(), bytesDeleted, blocksDeleted, bytesInTrash-bytesDeleted, blocksInTrash-blocksDeleted)
+ v.logger.Infof("EmptyTrash stats: Deleted %v bytes in %v blocks. Remaining in trash: %v bytes in %v blocks.", bytesDeleted, blocksDeleted, bytesInTrash-bytesDeleted, blocksInTrash-blocksDeleted)
}
type unixStats struct {
projects / arvados.git / blobdiff