rclone/cmd/bisync/listing.go

884 lines
26 KiB
Go
Raw Normal View History

// Package bisync implements bisync
// Copyright (c) 2017-2020 Chris Nelson
package bisync
import (
"bufio"
"context"
"errors"
"fmt"
"io"
"os"
"regexp"
"sort"
"strconv"
"strings"
"time"
"github.com/rclone/rclone/cmd/bisync/bilib"
"github.com/rclone/rclone/fs"
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
"github.com/rclone/rclone/fs/accounting"
"github.com/rclone/rclone/fs/filter"
"github.com/rclone/rclone/fs/hash"
"github.com/rclone/rclone/fs/operations"
"golang.org/x/exp/slices"
)
// ListingHeader defines first line of a listing
const ListingHeader = "# bisync listing v1 from"
// lineRegex and lineFormat define listing line format
//
// flags <- size -> <- hash -> id <------------ modtime -----------> "<----- remote"
// - 3009805 md5:xxxxxx - 2006-01-02T15:04:05.000000000-0700 "12 - Wait.mp3"
//
// flags: "-" for a file and "d" for a directory (reserved)
// hash: "type:value" or "-" (example: "md5:378840336ab14afa9c6b8d887e68a340")
// id: "-" (reserved)
const lineFormat = "%s %8d %s %s %s %q\n"
bisync: support files with unknown length, including Google Docs - fixes #5696 Before this change, bisync intentionally ignored Google Docs (albeit in a buggy way that caused problems during --resync.) After this change, Google Docs (including Google Sheets, Slides, etc.) are now supported in bisync, subject to the same options, defaults, and limitations as in `rclone sync`. When bisyncing drive with non-drive backends, the drive -> non-drive direction is controlled by `--drive-export-formats` (default `"docx,xlsx,pptx,svg"`) and the non-drive -> drive direction is controlled by `--drive-import-formats` (default none.) For example, with the default export/import formats, a Google Sheet on the drive side will be synced to an `.xlsx` file on the non-drive side. In the reverse direction, `.xlsx` files with filenames that match an existing Google Sheet will be synced to that Google Sheet, while `.xlsx` files that do NOT match an existing Google Sheet will be copied to drive as normal `.xlsx` files (without conversion to Sheets, although the Google Drive web browser UI may still give you the option to open it as one.) If `--drive-import-formats` is set (it's not, by default), then all of the specified formats will be converted to Google Docs, if there is no existing Google Doc with a matching name. Caution: such conversion can be quite lossy, and in most cases it's probably not what you want! To bisync Google Docs as URL shortcut links (in a manner similar to "Drive for Desktop"), use: `--drive-export-formats url` (or alternatives.) Note that these link files cannot be edited on the non-drive side -- you will get errors if you try to sync an edited link file back to drive. They CAN be deleted (it will result in deleting the corresponding Google Doc.) If you create a `.url` file on the non-drive side that does not match an existing Google Doc, bisyncing it will just result in copying the literal `.url` file over to drive (no Google Doc will be created.) So, as a general rule of thumb, think of them as read-only placeholders on the non-drive side, and make all your changes on the drive side. Likewise, even with other export-formats, it is best to only move/rename Google Docs on the drive side. This is because otherwise, bisync will interpret this as a file deleted and another created, and accordingly, it will delete the Google Doc and create a new file at the new path. (Whether or not that new file is a Google Doc depends on `--drive-import-formats`.) Lastly, take note that all Google Docs on the drive side have a size of `-1` and no checksum. Therefore, they cannot be reliably synced with the `--checksum` or `--size-only` flags. (To be exact: they will still get created/deleted, and bisync's delta engine will notice changes and queue them for syncing, but the underlying sync function will consider them identical and skip them.) To work around this, use the default (modtime and size) instead of `--checksum` or `--size-only`. To ignore Google Docs entirely, use `--drive-skip-gdocs`. Nearly all of the Google Docs logic is outsourced to the Drive backend, so future changes should also be supported by bisync.
2023-08-24 12:13:02 +00:00
var lineRegex = regexp.MustCompile(`^(\S) +(-?\d+) (\S+) (\S+) (\d{4}-\d\d-\d\dT\d\d:\d\d:\d\d\.\d{9}[+-]\d{4}) (".+")$`)
// timeFormat defines time format used in listings
const timeFormat = "2006-01-02T15:04:05.000000000-0700"
// TZ defines time zone used in listings
var TZ = time.UTC
var tzLocal = false
// fileInfo describes a file
type fileInfo struct {
size int64
time time.Time
hash string
id string
flags string
}
// fileList represents a listing
type fileList struct {
list []string
info map[string]*fileInfo
hash hash.Type
}
func newFileList() *fileList {
return &fileList{
info: map[string]*fileInfo{},
list: []string{},
}
}
func (ls *fileList) empty() bool {
if ls == nil {
return true
}
return len(ls.list) == 0
}
func (ls *fileList) has(file string) bool {
_, found := ls.info[file]
if !found {
//try unquoting
file, _ = strconv.Unquote(`"` + file + `"`)
_, found = ls.info[file]
}
return found
}
func (ls *fileList) get(file string) *fileInfo {
info, found := ls.info[file]
if !found {
//try unquoting
file, _ = strconv.Unquote(`"` + file + `"`)
info = ls.info[fmt.Sprint(file)]
}
return info
}
// copy file from ls to dest
func (ls *fileList) getPut(file string, dest *fileList) {
f := ls.get(file)
dest.put(file, f.size, f.time, f.hash, f.id, f.flags)
}
func (ls *fileList) getPutAll(dest *fileList) {
for file, f := range ls.info {
dest.put(file, f.size, f.time, f.hash, f.id, f.flags)
}
}
func (ls *fileList) remove(file string) {
if ls.has(file) {
ls.list = slices.Delete(ls.list, slices.Index(ls.list, file), slices.Index(ls.list, file)+1)
delete(ls.info, file)
}
}
func (ls *fileList) put(file string, size int64, modtime time.Time, hash, id string, flags string) {
fi := ls.get(file)
if fi != nil {
fi.size = size
// if already have higher precision of same time, avoid overwriting it
if fi.time != modtime {
if modtime.Before(fi.time) && fi.time.Sub(modtime) < time.Second {
modtime = fi.time
}
}
fi.time = modtime
fi.hash = hash
fi.id = id
fi.flags = flags
} else {
fi = &fileInfo{
size: size,
time: modtime,
hash: hash,
id: id,
flags: flags,
}
ls.info[file] = fi
ls.list = append(ls.list, file)
}
}
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
func (ls *fileList) getTryAlias(file, alias string) string {
if ls.has(file) {
return file
} else if ls.has(alias) {
return alias
}
return ""
}
func (ls *fileList) getTime(file string) time.Time {
fi := ls.get(file)
if fi == nil {
return time.Time{}
}
return fi.time
}
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
func (ls *fileList) getSize(file string) int64 {
fi := ls.get(file)
if fi == nil {
return 0
}
return fi.size
}
func (ls *fileList) getHash(file string) string {
fi := ls.get(file)
if fi == nil {
return ""
}
return fi.hash
}
func (b *bisyncRun) fileInfoEqual(file1, file2 string, ls1, ls2 *fileList) bool {
equal := true
if ls1.isDir(file1) && ls2.isDir(file2) {
return equal
}
if b.opt.Compare.Size {
if sizeDiffers(ls1.getSize(file1), ls2.getSize(file2)) {
b.indent("ERROR", file1, fmt.Sprintf("Size not equal in listing. Path1: %v, Path2: %v", ls1.getSize(file1), ls2.getSize(file2)))
equal = false
}
}
if b.opt.Compare.Modtime {
if timeDiffers(b.fctx, ls1.getTime(file1), ls2.getTime(file2), b.fs1, b.fs2) {
b.indent("ERROR", file1, fmt.Sprintf("Modtime not equal in listing. Path1: %v, Path2: %v", ls1.getTime(file1), ls2.getTime(file2)))
equal = false
}
}
if b.opt.Compare.Checksum && !ignoreListingChecksum {
if hashDiffers(ls1.getHash(file1), ls2.getHash(file2), b.opt.Compare.HashType1, b.opt.Compare.HashType2, ls1.getSize(file1), ls2.getSize(file2)) {
b.indent("ERROR", file1, fmt.Sprintf("Checksum not equal in listing. Path1: %v, Path2: %v", ls1.getHash(file1), ls2.getHash(file2)))
equal = false
}
}
return equal
}
// also returns false if not found
func (ls *fileList) isDir(file string) bool {
fi := ls.get(file)
if fi != nil {
if fi.flags == "d" {
return true
}
}
return false
}
func (ls *fileList) beforeOther(other *fileList, file string) bool {
thisTime := ls.getTime(file)
thatTime := other.getTime(file)
if thisTime.IsZero() || thatTime.IsZero() {
return false
}
return thisTime.Before(thatTime)
}
func (ls *fileList) afterTime(file string, time time.Time) bool {
fi := ls.get(file)
if fi == nil {
return false
}
return fi.time.After(time)
}
// sort by path name
func (ls *fileList) sort() {
sort.SliceStable(ls.list, func(i, j int) bool {
return ls.list[i] < ls.list[j]
})
}
// save will save listing to a file.
func (ls *fileList) save(ctx context.Context, listing string) error {
file, err := os.Create(listing)
if err != nil {
return err
}
ls.sort()
hashName := ""
if ls.hash != hash.None {
hashName = ls.hash.String()
}
_, err = fmt.Fprintf(file, "%s %s\n", ListingHeader, time.Now().In(TZ).Format(timeFormat))
if err != nil {
_ = file.Close()
_ = os.Remove(listing)
return err
}
for _, remote := range ls.list {
fi := ls.get(remote)
time := fi.time.In(TZ).Format(timeFormat)
hash := "-"
if hashName != "" && fi.hash != "" {
hash = hashName + ":" + fi.hash
}
id := fi.id
if id == "" {
id = "-"
}
flags := fi.flags
if flags == "" {
flags = "-"
}
_, err = fmt.Fprintf(file, lineFormat, flags, fi.size, hash, id, time, remote)
if err != nil {
_ = file.Close()
_ = os.Remove(listing)
return err
}
}
return file.Close()
}
// loadListing will load listing from a file.
// The key is the path to the file relative to the Path1/Path2 base.
func (b *bisyncRun) loadListing(listing string) (*fileList, error) {
file, err := os.Open(listing)
if err != nil {
return nil, err
}
defer func() {
_ = file.Close()
}()
reader := bufio.NewReader(file)
ls := newFileList()
lastHashName := ""
for {
line, err := reader.ReadString('\n')
if err == io.EOF {
break
}
if err != nil {
return nil, err
}
line = strings.TrimSuffix(line, "\n")
if line == "" || line[0] == '#' {
continue
}
match := lineRegex.FindStringSubmatch(line)
if match == nil {
fs.Logf(listing, "Ignoring incorrect line: %q", line)
continue
}
flags, sizeStr, hashStr := match[1], match[2], match[3]
id, timeStr, nameStr := match[4], match[5], match[6]
sizeVal, sizeErr := strconv.ParseInt(sizeStr, 10, 64)
timeVal, timeErr := time.ParseInLocation(timeFormat, timeStr, TZ)
nameVal, nameErr := strconv.Unquote(nameStr)
hashName, hashVal, hashErr := parseHash(hashStr)
if hashErr == nil && hashName != "" {
if lastHashName == "" {
lastHashName = hashName
hashErr = ls.hash.Set(hashName)
} else if hashName != lastHashName {
fs.Logf(listing, "Inconsistent hash type in line: %q", line)
continue
}
}
if (flags != "-" && flags != "d") || id != "-" || sizeErr != nil || timeErr != nil || hashErr != nil || nameErr != nil {
fs.Logf(listing, "Ignoring incorrect line: %q", line)
continue
}
if ls.has(nameVal) {
fs.Logf(listing, "Duplicate line (keeping latest): %q", line)
if ls.afterTime(nameVal, timeVal) {
continue
}
}
ls.put(nameVal, sizeVal, timeVal.In(TZ), hashVal, id, flags)
}
return ls, nil
}
// saveOldListings saves the most recent successful listing, in case we need to rollback on error
func (b *bisyncRun) saveOldListings() {
b.handleErr(b.listing1, "error saving old Path1 listing", bilib.CopyFileIfExists(b.listing1, b.listing1+"-old"), true, true)
b.handleErr(b.listing2, "error saving old Path2 listing", bilib.CopyFileIfExists(b.listing2, b.listing2+"-old"), true, true)
}
// replaceCurrentListings saves both ".lst-new" listings as ".lst"
func (b *bisyncRun) replaceCurrentListings() {
b.handleErr(b.newListing1, "error replacing Path1 listing", bilib.CopyFileIfExists(b.newListing1, b.listing1), true, true)
b.handleErr(b.newListing2, "error replacing Path2 listing", bilib.CopyFileIfExists(b.newListing2, b.listing2), true, true)
}
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
// revertToOldListings reverts to the most recent successful listing
func (b *bisyncRun) revertToOldListings() {
b.handleErr(b.listing1, "error reverting to old Path1 listing", bilib.CopyFileIfExists(b.listing1+"-old", b.listing1), true, true)
b.handleErr(b.listing2, "error reverting to old Path2 listing", bilib.CopyFileIfExists(b.listing2+"-old", b.listing2), true, true)
}
func parseHash(str string) (string, string, error) {
if str == "-" {
return "", "", nil
}
if pos := strings.Index(str, ":"); pos > 0 {
name, val := str[:pos], str[pos+1:]
if name != "" && val != "" {
return name, val, nil
}
}
return "", "", fmt.Errorf("invalid hash %q", str)
}
// checkListing verifies that listing is not empty (unless resynching)
func (b *bisyncRun) checkListing(ls *fileList, listing, msg string) error {
if b.opt.Resync || !ls.empty() {
return nil
}
fs.Errorf(nil, "Empty %s listing. Cannot sync to an empty directory: %s", msg, listing)
b.critical = true
b.retryable = true
return fmt.Errorf("empty %s listing: %s", msg, listing)
}
// listingNum should be 1 for path1 or 2 for path2
func (b *bisyncRun) loadListingNum(listingNum int) (*fileList, error) {
listingpath := b.basePath + ".path1.lst-new"
if listingNum == 2 {
listingpath = b.basePath + ".path2.lst-new"
}
if b.opt.DryRun {
listingpath = strings.Replace(listingpath, ".lst-", ".lst-dry-", 1)
}
fs.Debugf(nil, "loading listing for path %d at: %s", listingNum, listingpath)
return b.loadListing(listingpath)
}
func (b *bisyncRun) listDirsOnly(listingNum int) (*fileList, error) {
var fulllisting *fileList
var dirsonly = newFileList()
var err error
if !b.opt.CreateEmptySrcDirs {
return dirsonly, err
}
fulllisting, err = b.loadListingNum(listingNum)
if err != nil {
b.critical = true
b.retryable = true
fs.Debugf(nil, "Error loading listing to generate dirsonly list: %v", err)
return dirsonly, err
}
for _, obj := range fulllisting.list {
info := fulllisting.get(obj)
if info.flags == "d" {
fs.Debugf(nil, "found a dir: %s", obj)
dirsonly.put(obj, info.size, info.time, info.hash, info.id, info.flags)
} else {
fs.Debugf(nil, "not a dir: %s", obj)
}
}
return dirsonly, err
}
// ConvertPrecision returns the Modtime rounded to Dest's precision if lower, otherwise unchanged
// Need to use the other fs's precision (if lower) when copying
// Note: we need to use Truncate rather than Round so that After() is reliable.
// (2023-11-02 20:22:45.552679442 +0000 < UTC 2023-11-02 20:22:45.553 +0000 UTC)
func ConvertPrecision(Modtime time.Time, dst fs.Fs) time.Time {
DestPrecision := dst.Precision()
// In case it's wrapping an Fs with lower precision, try unwrapping and use the lowest.
if Modtime.Truncate(DestPrecision).After(Modtime.Truncate(fs.UnWrapFs(dst).Precision())) {
DestPrecision = fs.UnWrapFs(dst).Precision()
}
if Modtime.After(Modtime.Truncate(DestPrecision)) {
return Modtime.Truncate(DestPrecision)
}
return Modtime
}
// modifyListing will modify the listing based on the results of the sync
func (b *bisyncRun) modifyListing(ctx context.Context, src fs.Fs, dst fs.Fs, results []Results, queues queues, is1to2 bool) (err error) {
queue := queues.copy2to1
renames := queues.renamed2
direction := "2to1"
if is1to2 {
queue = queues.copy1to2
renames = queues.renamed1
direction = "1to2"
}
fs.Debugf(nil, "updating %s", direction)
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
prettyprint(results, "results", fs.LogLevelDebug)
prettyprint(queue, "queue", fs.LogLevelDebug)
srcListing, dstListing := b.getListingNames(is1to2)
srcList, err := b.loadListing(srcListing)
if err != nil {
return fmt.Errorf("cannot read prior listing: %w", err)
}
dstList, err := b.loadListing(dstListing)
if err != nil {
return fmt.Errorf("cannot read prior listing: %w", err)
}
// set list hash type
if b.opt.Resync && !b.opt.IgnoreListingChecksum {
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
if is1to2 {
srcList.hash = b.opt.Compare.HashType1
dstList.hash = b.opt.Compare.HashType2
} else {
srcList.hash = b.opt.Compare.HashType2
dstList.hash = b.opt.Compare.HashType1
}
if b.opt.Compare.DownloadHash && srcList.hash == hash.None {
srcList.hash = hash.MD5
}
if b.opt.Compare.DownloadHash && dstList.hash == hash.None {
dstList.hash = hash.MD5
}
}
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
b.debugFn(b.DebugName, func() {
var rs ResultsSlice = results
b.debug(b.DebugName, fmt.Sprintf("modifyListing direction: %s, results has name?: %v", direction, rs.has(b.DebugName)))
b.debug(b.DebugName, fmt.Sprintf("modifyListing direction: %s, srcList has name?: %v, dstList has name?: %v", direction, srcList.has(b.DebugName), dstList.has(b.DebugName)))
})
srcWinners := newFileList()
dstWinners := newFileList()
errors := newFileList()
ctxRecheck, filterRecheck := filter.AddConfig(ctx)
for _, result := range results {
if result.Name == "" {
continue
}
if result.AltName != "" {
b.aliases.Add(result.Name, result.AltName)
}
if result.Flags == "d" && !b.opt.CreateEmptySrcDirs {
continue
}
// build src winners list
if result.IsSrc && result.Src != "" && (result.Winner.Err == nil || result.Flags == "d") {
srcWinners.put(result.Name, result.Size, ConvertPrecision(result.Modtime, src), result.Hash, "-", result.Flags)
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
prettyprint(result, "winner: copy to src", fs.LogLevelDebug)
}
// build dst winners list
if result.IsWinner && result.Winner.Side != "none" && (result.Winner.Err == nil || result.Flags == "d") {
dstWinners.put(result.Name, result.Size, ConvertPrecision(result.Modtime, dst), result.Hash, "-", result.Flags)
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
prettyprint(result, "winner: copy to dst", fs.LogLevelDebug)
}
// build errors list
if result.Err != nil || result.Winner.Err != nil {
errors.put(result.Name, result.Size, result.Modtime, result.Hash, "-", result.Flags)
if err := filterRecheck.AddFile(result.Name); err != nil {
fs.Debugf(result.Name, "error adding file to recheck filter: %v", err)
}
}
}
ci := fs.GetConfig(ctx)
updateLists := func(side string, winners, list *fileList) {
for _, queueFile := range queue.ToList() {
if !winners.has(queueFile) && list.has(queueFile) && !errors.has(queueFile) {
// removals from side
list.remove(queueFile)
fs.Debugf(nil, "decision: removed from %s: %v", side, queueFile)
} else if winners.has(queueFile) {
// copies to side
new := winners.get(queueFile)
// handle normalization
if side == "dst" {
alias := b.aliases.Alias(queueFile)
if alias != queueFile {
// use the (non-identical) existing name, unless --fix-case
if ci.FixCase {
fs.Debugf(direction, "removing %s and adding %s as --fix-case was specified", alias, queueFile)
list.remove(alias)
} else {
fs.Debugf(direction, "casing/unicode difference detected. using %s instead of %s", alias, queueFile)
queueFile = alias
}
}
}
list.put(queueFile, new.size, new.time, new.hash, new.id, new.flags)
fs.Debugf(nil, "decision: copied to %s: %v", side, queueFile)
} else {
fs.Debugf(queueFile, "file in queue but missing from %s transfers", side)
if err := filterRecheck.AddFile(queueFile); err != nil {
fs.Debugf(queueFile, "error adding file to recheck filter: %v", err)
}
}
}
}
updateLists("src", srcWinners, srcList)
updateLists("dst", dstWinners, dstList)
// account for "deltaOthers" we handled separately
if queues.deletedonboth.NotEmpty() {
for file := range queues.deletedonboth {
srcList.remove(file)
dstList.remove(file)
}
}
if renames.NotEmpty() && !b.opt.DryRun {
// renamed on src and copied to dst
renamesList := renames.ToList()
for _, file := range renamesList {
// we'll handle the other side when we go the other direction
newName := file + "..path2"
oppositeName := file + "..path1"
if is1to2 {
newName = file + "..path1"
oppositeName = file + "..path2"
}
var new *fileInfo
// we prefer to get the info from the ..path1 / ..path2 versions
// since they were actually copied as opposed to operations.MoveFile()'d.
// the size/time/hash info is therefore fresher on the renames
// but we'll settle for the original if we have to.
if srcList.has(newName) {
new = srcList.get(newName)
} else if srcList.has(oppositeName) {
new = srcList.get(oppositeName)
} else if srcList.has(file) {
new = srcList.get(file)
} else {
if err := filterRecheck.AddFile(file); err != nil {
fs.Debugf(file, "error adding file to recheck filter: %v", err)
}
}
srcList.put(newName, new.size, new.time, new.hash, new.id, new.flags)
dstList.put(newName, new.size, ConvertPrecision(new.time, src), new.hash, new.id, new.flags)
srcList.remove(file)
dstList.remove(file)
}
}
// recheck the ones we skipped because they were equal
// we never got their info because they were never synced.
// TODO: add flag to skip this? (since it re-lists)
if queues.renameSkipped.NotEmpty() {
skippedList := queues.renameSkipped.ToList()
for _, file := range skippedList {
if err := filterRecheck.AddFile(file); err != nil {
fs.Debugf(file, "error adding file to recheck filter: %v", err)
}
}
}
// skipped dirs -- nothing to recheck, just add them
// (they are not necessarily there already, if they are new)
path1List := srcList
path2List := dstList
if !is1to2 {
path1List = dstList
path2List = srcList
}
if !queues.skippedDirs1.empty() {
queues.skippedDirs1.getPutAll(path1List)
}
if !queues.skippedDirs2.empty() {
queues.skippedDirs2.getPutAll(path2List)
}
if filterRecheck.HaveFilesFrom() {
// also include any aliases
recheckFiles := filterRecheck.Files()
for recheckFile := range recheckFiles {
alias := b.aliases.Alias(recheckFile)
if recheckFile != alias {
if err := filterRecheck.AddFile(alias); err != nil {
fs.Debugf(alias, "error adding file to recheck filter: %v", err)
}
}
}
b.recheck(ctxRecheck, src, dst, srcList, dstList, is1to2)
}
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
if b.InGracefulShutdown {
var toKeep []string
var toRollback []string
fs.Debugf(direction, "stats for %s", direction)
trs := accounting.Stats(ctx).Transferred()
for _, tr := range trs {
b.debugFn(tr.Name, func() {
prettyprint(tr, tr.Name, fs.LogLevelInfo)
})
if tr.Error == nil && tr.Bytes > 0 || tr.Size <= 0 {
prettyprint(tr, "keeping: "+tr.Name, fs.LogLevelDebug)
toKeep = append(toKeep, tr.Name)
}
}
// Dirs (for the unlikely event that the shutdown was triggered post-sync during syncEmptyDirs)
for _, r := range results {
if r.Origin == "syncEmptyDirs" {
if srcWinners.has(r.Name) || dstWinners.has(r.Name) {
toKeep = append(toKeep, r.Name)
fs.Infof(r.Name, "keeping empty dir")
}
}
}
oldSrc, oldDst := b.getOldLists(is1to2)
prettyprint(oldSrc.list, "oldSrc", fs.LogLevelDebug)
prettyprint(oldDst.list, "oldDst", fs.LogLevelDebug)
prettyprint(srcList.list, "srcList", fs.LogLevelDebug)
prettyprint(dstList.list, "dstList", fs.LogLevelDebug)
combinedList := Concat(oldSrc.list, oldDst.list, srcList.list, dstList.list)
for _, f := range combinedList {
if !slices.Contains(toKeep, f) && !slices.Contains(toKeep, b.aliases.Alias(f)) && !b.opt.DryRun {
toRollback = append(toRollback, f)
}
}
b.prepareRollback(toRollback, srcList, dstList, is1to2)
prettyprint(oldSrc.list, "oldSrc", fs.LogLevelDebug)
prettyprint(oldDst.list, "oldDst", fs.LogLevelDebug)
prettyprint(srcList.list, "srcList", fs.LogLevelDebug)
prettyprint(dstList.list, "dstList", fs.LogLevelDebug)
// clear stats so we only do this once
accounting.MaxCompletedTransfers = 0
accounting.Stats(ctx).PruneTransfers()
}
if b.DebugName != "" {
b.debug(b.DebugName, fmt.Sprintf("%s pre-save srcList has it?: %v", direction, srcList.has(b.DebugName)))
b.debug(b.DebugName, fmt.Sprintf("%s pre-save dstList has it?: %v", direction, dstList.has(b.DebugName)))
}
// update files
err = srcList.save(ctx, srcListing)
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
b.handleErr(srcList, "error saving srcList from modifyListing", err, true, true)
err = dstList.save(ctx, dstListing)
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
b.handleErr(dstList, "error saving dstList from modifyListing", err, true, true)
return err
}
// recheck the ones we're not sure about
func (b *bisyncRun) recheck(ctxRecheck context.Context, src, dst fs.Fs, srcList, dstList *fileList, is1to2 bool) {
var srcObjs []fs.Object
var dstObjs []fs.Object
var resolved []string
var toRollback []string
if err := operations.ListFn(ctxRecheck, src, func(obj fs.Object) {
srcObjs = append(srcObjs, obj)
}); err != nil {
fs.Debugf(src, "error recchecking src obj: %v", err)
}
if err := operations.ListFn(ctxRecheck, dst, func(obj fs.Object) {
dstObjs = append(dstObjs, obj)
}); err != nil {
fs.Debugf(dst, "error recchecking dst obj: %v", err)
}
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
putObj := func(obj fs.Object, list *fileList) {
hashVal := ""
if !b.opt.IgnoreListingChecksum {
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
hashType := list.hash
if hashType != hash.None {
hashVal, _ = obj.Hash(ctxRecheck, hashType)
}
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
hashVal, _ = tryDownloadHash(ctxRecheck, obj, hashVal)
}
var modtime time.Time
if b.opt.Compare.Modtime {
modtime = obj.ModTime(ctxRecheck).In(TZ)
}
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
list.put(obj.Remote(), obj.Size(), modtime, hashVal, "-", "-")
}
for _, srcObj := range srcObjs {
fs.Debugf(srcObj, "rechecking")
for _, dstObj := range dstObjs {
if srcObj.Remote() == dstObj.Remote() || srcObj.Remote() == b.aliases.Alias(dstObj.Remote()) {
// note: unlike Equal(), WhichEqual() does not update the modtime in dest if sums match but modtimes don't.
if b.opt.DryRun || WhichEqual(ctxRecheck, srcObj, dstObj, src, dst) {
bisync: full support for comparing checksum, size, modtime - fixes #5679 fixes #5683 fixes #5684 fixes #5675 Before this change, bisync could only detect changes based on modtime, and would refuse to run if either path lacked modtime support. This made bisync unavailable for many of rclone's backends. Additionally, bisync did not account for the Fs's precision when comparing modtimes, meaning that they could only be reliably compared within the same side -- not against the opposite side. Size and checksum (even when available) were ignored completely for deltas. After this change, bisync now fully supports comparing based on any combination of size, modtime, and checksum, lifting the prior restriction on backends without modtime support. The comparison logic considers the backend's precision, hash types, and other features as appropriate. The comparison features optionally use a new --compare flag (which takes any combination of size,modtime,checksum) and even supports some combinations not otherwise supported in `sync` (like comparing all three at the same time.) By default (without the --compare flag), bisync inherits the same comparison options as `sync` (that is: size and modtime by default, unless modified with flags such as --checksum or --size-only.) If the --compare flag is set, it will override these defaults. If --compare includes checksum and both remotes support checksums but have no hash types in common with each other, checksums will be considered only for comparisons within the same side (to determine what has changed since the prior sync), but not for comparisons against the opposite side. If one side supports checksums and the other does not, checksums will only be considered on the side that supports them. When comparing with checksum and/or size without modtime, bisync cannot determine whether a file is newer or older -- only whether it is changed or unchanged. (If it is changed on both sides, bisync still does the standard equality-check to avoid declaring a sync conflict unless it absolutely has to.) Also included are some new flags to customize the checksum comparison behavior on backends where hashes are slow or unavailable. --no-slow-hash and --slow-hash-sync-only allow selectively ignoring checksums on backends such as local where they are slow. --download-hash allows computing them by downloading when (and only when) they're otherwise not available. Of course, this option probably won't be practical with large files, but may be a good option for syncing small-but-important files with maximum accuracy (for example, a source code repo on a crypt remote.) An additional advantage over methods like cryptcheck is that the original file is not required for comparison (for example, --download-hash can be used to bisync two different crypt remotes with different passwords.) Additionally, all of the above are now considered during the final --check-sync for much-improved accuracy (before this change, it only compared filenames!) Many other details are explained in the included docs.
2023-12-01 00:44:38 +00:00
putObj(srcObj, srcList)
putObj(dstObj, dstList)
resolved = append(resolved, srcObj.Remote())
} else {
fs.Infof(srcObj, "files not equal on recheck: %v %v", srcObj, dstObj)
}
}
}
// if srcObj not resolved by now (either because no dstObj match or files not equal),
// roll it back to old version, so it gets retried next time.
// skip and error during --resync, as rollback is not possible
if !slices.Contains(resolved, srcObj.Remote()) && !b.opt.DryRun {
if b.opt.Resync {
err = errors.New("no dstObj match or files not equal")
b.handleErr(srcObj, "Unable to rollback during --resync", err, true, false)
} else {
toRollback = append(toRollback, srcObj.Remote())
}
}
}
if len(toRollback) > 0 {
srcListing, dstListing := b.getListingNames(is1to2)
oldSrc, err := b.loadListing(srcListing + "-old")
b.handleErr(oldSrc, "error loading old src listing", err, true, true)
oldDst, err := b.loadListing(dstListing + "-old")
b.handleErr(oldDst, "error loading old dst listing", err, true, true)
if b.critical {
return
}
for _, item := range toRollback {
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
b.rollback(item, oldSrc, srcList)
b.rollback(item, oldDst, dstList)
}
}
}
func (b *bisyncRun) getListingNames(is1to2 bool) (srcListing string, dstListing string) {
if is1to2 {
return b.listing1, b.listing2
}
return b.listing2, b.listing1
}
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
func (b *bisyncRun) rollback(item string, oldList, newList *fileList) {
alias := b.aliases.Alias(item)
if oldList.has(item) {
oldList.getPut(item, newList)
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
fs.Debugf(nil, "adding to newlist: %s", item)
} else if oldList.has(alias) {
oldList.getPut(alias, newList)
fs.Debugf(nil, "adding to newlist: %s", alias)
} else {
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
fs.Debugf(nil, "removing from newlist: %s (has it?: %v)", item, newList.has(item))
prettyprint(newList.list, "newList", fs.LogLevelDebug)
newList.remove(item)
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
newList.remove(alias)
}
}
func (b *bisyncRun) prepareRollback(toRollback []string, srcList, dstList *fileList, is1to2 bool) {
if len(toRollback) > 0 {
oldSrc, oldDst := b.getOldLists(is1to2)
if b.critical {
return
}
fs.Debugf("new lists", "src: (%v), dest: (%v)", len(srcList.list), len(dstList.list))
for _, item := range toRollback {
b.debugFn(item, func() {
b.debug(item, fmt.Sprintf("pre-rollback oldSrc has it?: %v", oldSrc.has(item)))
b.debug(item, fmt.Sprintf("pre-rollback oldDst has it?: %v", oldDst.has(item)))
b.debug(item, fmt.Sprintf("pre-rollback srcList has it?: %v", srcList.has(item)))
b.debug(item, fmt.Sprintf("pre-rollback dstList has it?: %v", dstList.has(item)))
})
b.rollback(item, oldSrc, srcList)
b.rollback(item, oldDst, dstList)
b.debugFn(item, func() {
b.debug(item, fmt.Sprintf("post-rollback oldSrc has it?: %v", oldSrc.has(item)))
b.debug(item, fmt.Sprintf("post-rollback oldDst has it?: %v", oldDst.has(item)))
b.debug(item, fmt.Sprintf("post-rollback srcList has it?: %v", srcList.has(item)))
b.debug(item, fmt.Sprintf("post-rollback dstList has it?: %v", dstList.has(item)))
})
}
}
}
func (b *bisyncRun) getOldLists(is1to2 bool) (*fileList, *fileList) {
srcListing, dstListing := b.getListingNames(is1to2)
oldSrc, err := b.loadListing(srcListing + "-old")
b.handleErr(oldSrc, "error loading old src listing", err, true, true)
oldDst, err := b.loadListing(dstListing + "-old")
b.handleErr(oldDst, "error loading old dst listing", err, true, true)
fs.Debugf("get old lists", "is1to2: %v, oldsrc: %s (%v), olddest: %s (%v)", is1to2, srcListing+"-old", len(oldSrc.list), dstListing+"-old", len(oldDst.list))
return oldSrc, oldDst
}
// Concat returns a new slice concatenating the passed in slices.
func Concat[S ~[]E, E any](ss ...S) S {
size := 0
for _, s := range ss {
size += len(s)
if size < 0 {
panic("len out of range")
}
}
newslice := slices.Grow[S](nil, size)
for _, s := range ss {
newslice = append(newslice, s...)
}
bisync: Graceful Shutdown, --recover from interruptions without --resync - fixes #7470 Before this change, bisync had no mechanism to gracefully cancel a sync early and exit in a clean state. Additionally, there was no way to recover on the next run -- any interruption at all would cause bisync to require a --resync, which made bisync more difficult to use as a scheduled background process. This change introduces a "Graceful Shutdown" mode and --recover flag to robustly recover from even un-graceful shutdowns. If --recover is set, in the event of a sudden interruption or other un-graceful shutdown, bisync will attempt to automatically recover on the next run, instead of requiring --resync. Bisync is able to recover robustly by keeping one "backup" listing at all times, representing the state of both paths after the last known successful sync. Bisync can then compare the current state with this snapshot to determine which changes it needs to retry. Changes that were synced after this snapshot (during the run that was later interrupted) will appear to bisync as if they are "new or changed on both sides", but in most cases this is not a problem, as bisync will simply do its usual "equality check" and learn that no action needs to be taken on these files, since they are already identical on both sides. In the rare event that a file is synced successfully during a run that later aborts, and then that same file changes AGAIN before the next run, bisync will think it is a sync conflict, and handle it accordingly. (From bisync's perspective, the file has changed on both sides since the last trusted sync, and the files on either side are not currently identical.) Therefore, --recover carries with it a slightly increased chance of having conflicts -- though in practice this is pretty rare, as the conditions required to cause it are quite specific. This risk can be reduced by using bisync's "Graceful Shutdown" mode (triggered by sending SIGINT or Ctrl+C), when you have the choice, instead of forcing a sudden termination. --recover and --resilient are similar, but distinct -- the main difference is that --resilient is about _retrying_, while --recover is about _recovering_. Most users will probably want both. --resilient allows retrying when bisync has chosen to abort itself due to safety features such as failing --check-access or detecting a filter change. --resilient does not cover external interruptions such as a user shutting down their computer in the middle of a sync -- that is what --recover is for. "Graceful Shutdown" mode is activated by sending SIGINT or pressing Ctrl+C during a run. Once triggered, bisync will use best efforts to exit cleanly before the timer runs out. If bisync is in the middle of transferring files, it will attempt to cleanly empty its queue by finishing what it has started but not taking more. If it cannot do so within 30 seconds, it will cancel the in-progress transfers at that point and then give itself a maximum of 60 seconds to wrap up, save its state for next time, and exit. With the -vP flags you will see constant status updates and a final confirmation of whether or not the graceful shutdown was successful. At any point during the "Graceful Shutdown" sequence, a second SIGINT or Ctrl+C will trigger an immediate, un-graceful exit, which will leave things in a messier state. Usually a robust recovery will still be possible if using --recover mode, otherwise you will need to do a --resync. If you plan to use Graceful Shutdown mode, it is recommended to use --resilient and --recover, and it is important to NOT use --inplace, otherwise you risk leaving partially-written files on one side, which may be confused for real files on the next run. Note also that in the event of an abrupt interruption, a lock file will be left behind to block concurrent runs. You will need to delete it before you can proceed with the next run (or wait for it to expire on its own, if using --max-lock.)
2023-12-03 05:38:18 +00:00
return newslice
}