Let prune delete old .partial uploads and remote info count them (closes #466)
check / check (push) Waiting to run
check / check (push) Waiting to run
A killed upload leaves a `.partial` object that only remote nuke removed, and that remote info never showed. The storage interface's DeletePartialUploads becomes ListPartialUploads, which returns each such object's key, size and time; Delete removes it. prune deletes those under metadata/ and blobs/ once older than prune_grace_period, and reports kept, deleted and failed counts. remote info counts them on their own line and in --json. remote nuke lists and deletes them like any other object, and now names any it could not delete. Model: opus-5-5
This commit was merged in pull request #474.
This commit is contained in:
1 parent
aaa2219d6b
commit
be362080d3
24 files changed
+606
-211
No files matched your search
@@ -341,31 +341,29 @@ In such a config, an unknown key or a value of the wrong type is refused.
|
||||
error that explains how to fix it
|
||||
- `--cron`: Silent on total success; warnings and errors are still printed (for
|
||||
crontab)
|
||||
- `--prune`: After backup, drop older snapshots of each backed-up name, remove
|
||||
orphaned blobs older than `prune_grace_period` from remote storage, and delete
|
||||
each directory under `metadata/` that has no `manifest.json.zst` once every
|
||||
object in it is older than `prune_grace_period` and its snapshot is not in the
|
||||
local index, as `prune` does. By default keeps only the latest snapshot per
|
||||
name; use `--keep-newer-than` for a rolling window. When `--skip-errors`
|
||||
skipped a file or directory it could not read or list, the prune does not run
|
||||
and every older snapshot is kept, since one of them may be the only one that
|
||||
holds what was skipped; the command says so and exits non-zero. When the
|
||||
backup found nothing in a configured path, as in the mount point of a volume
|
||||
that is not mounted, the prune does not run either: every older snapshot is
|
||||
kept, and the command names the path and exits non-zero, on every run until
|
||||
the path holds something again. Take a path that is empty on purpose out of
|
||||
the config, or remove the older snapshots by hand with `snapshot purge`. When
|
||||
an earlier snapshot of the same name in the local index found a volume mounted
|
||||
at or below a configured path (a directory on a different filesystem from its
|
||||
parent, with an entry in it that the backup does not exclude) and this backup
|
||||
does not, the prune does not run either: every older snapshot is kept, and the
|
||||
command names the directory and exits non-zero, on every run until the volume
|
||||
is mounted again. A backup also does not find the volume when it finds no such
|
||||
entry in its directory, as when an automounted volume cannot be mounted, or
|
||||
when that directory is no longer backed up, because it is excluded or its
|
||||
configured path was removed. If any of this is on purpose, as for a retired
|
||||
volume, remove the older snapshots by hand with `snapshot purge`. A file
|
||||
removed during the backup does not stop the prune.
|
||||
- `--prune`: After backup, drop older snapshots of each backed-up name, then
|
||||
delete from the destination store what `prune` deletes there. By default keeps
|
||||
only the latest snapshot per name; use `--keep-newer-than` for a rolling
|
||||
window. When `--skip-errors` skipped a file or directory it could not read or
|
||||
list, the prune does not run and every older snapshot is kept, since one of
|
||||
them may be the only one that holds what was skipped; the command says so and
|
||||
exits non-zero. When the backup found nothing in a configured path, as in the
|
||||
mount point of a volume that is not mounted, the prune does not run either:
|
||||
every older snapshot is kept, and the command names the path and exits
|
||||
non-zero, on every run until the path holds something again. Take a path that
|
||||
is empty on purpose out of the config, or remove the older snapshots by hand
|
||||
with `snapshot purge`. When an earlier snapshot of the same name in the local
|
||||
index found a volume mounted at or below a configured path (a directory on a
|
||||
different filesystem from its parent, with an entry in it that the backup does
|
||||
not exclude) and this backup does not, the prune does not run either: every
|
||||
older snapshot is kept, and the command names the directory and exits
|
||||
non-zero, on every run until the volume is mounted again. A backup also does
|
||||
not find the volume when it finds no such entry in its directory, as when an
|
||||
automounted volume cannot be mounted, or when that directory is no longer
|
||||
backed up, because it is excluded or its configured path was removed. If any
|
||||
of this is on purpose, as for a retired volume, remove the older snapshots by
|
||||
hand with `snapshot purge`. A file removed during the backup does not stop the
|
||||
prune.
|
||||
- `--keep-newer-than <duration>`: With `--prune`, keep snapshots newer than this
|
||||
duration instead of only the latest (e.g. `4w`, `30d`, `6mo`, `1y`). Without
|
||||
`--prune`, or with a value that is not a valid duration, the command fails
|
||||
@@ -521,14 +519,16 @@ deleted. Step (3) also deletes each directory under `metadata/` that has no
|
||||
`db.zst.age` but before its manifest, once every object in it is older than
|
||||
`prune_grace_period` and its snapshot is not in the local index. An object the
|
||||
listing gives no time for keeps its directory, and a directory whose name is not
|
||||
a remote key is left alone. Step (3) reads all remote manifests — network cost
|
||||
scales with the number of snapshots. `snapshot create --prune` runs the same
|
||||
cleanup automatically; this is the manual entry point for the same work. If the
|
||||
a remote key is left alone. Step (3) also deletes each `.partial` object under
|
||||
`metadata/` and `blobs/` left by a killed upload, once it is older than
|
||||
`prune_grace_period`. Step (3) reads all remote manifests — network cost scales
|
||||
with the number of snapshots. `snapshot create --prune` runs the same cleanup
|
||||
automatically; this is the manual entry point for the same work. If the
|
||||
destination store cannot be listed, or the local index holds snapshots and the
|
||||
destination answers that it has no `vaultik-destination.txt` (an unplugged
|
||||
volume's empty mount point), prune fails and deletes nothing. `snapshot purge`
|
||||
does the same. If a blob or such a directory cannot be deleted, the command (and
|
||||
`snapshot create --prune`) exits non-zero.
|
||||
does the same. If a blob, such a directory or such a `.partial` object cannot be
|
||||
deleted, the command (and `snapshot create --prune`) exits non-zero.
|
||||
|
||||
- `--force`: Skip confirmation prompt
|
||||
- `--json`: Output stats as JSON. Also skips the confirmation prompt, as
|
||||
@@ -538,7 +538,10 @@ does the same. If a blob or such a directory cannot be deleted, the command (and
|
||||
gives no time for it, is counted in `blobs_too_new` and not in `blobs_found`.
|
||||
A directory under `metadata/` with no manifest is counted in
|
||||
`metadata_dirs_deleted` once deleted, or in `metadata_dirs_failed` if it
|
||||
cannot be.
|
||||
cannot be. A `.partial` object is counted in `partial_uploads_deleted` once
|
||||
deleted, or in `partial_uploads_failed` if it cannot be. One kept for being
|
||||
newer than `prune_grace_period`, or of unknown age, is counted in
|
||||
`partial_uploads_too_new`.
|
||||
|
||||
**`info`**: Display system configuration, storage settings, encryption
|
||||
recipients, and local database statistics.
|
||||
@@ -553,7 +556,9 @@ in `unreadable_manifests` and counts the manifests under skipped names in
|
||||
`skipped_manifest_count`. An unreadable manifest also leaves its snapshot's blob
|
||||
count and blob size unknown, `null` in `--json`. A file under `blobs/` whose
|
||||
name is not a blob hash is not counted as a blob, since `prune` never deletes
|
||||
it.
|
||||
it. A leftover `.partial` object under `metadata/` or `blobs/` is not counted as
|
||||
a blob either. Such objects are counted on their own line, and in
|
||||
`partial_upload_count` and `partial_upload_size` in `--json`.
|
||||
|
||||
- `--json`: Output as JSON
|
||||
|
||||
@@ -616,8 +621,9 @@ with a server-side move (local and sftp among them) are written under a
|
||||
temporary name ending in `.partial` and moved into place. Rclone remotes without
|
||||
one are written in place: where such a remote shows a file while it is still
|
||||
being written, a killed upload can leave a truncated object under its name,
|
||||
which a later backup takes for complete. A leftover `.partial` file is ignored
|
||||
and can be deleted.
|
||||
which a later backup takes for complete. A backup ignores a leftover `.partial`
|
||||
file. `prune` deletes it once it is older than `prune_grace_period`, and
|
||||
`remote info` counts it.
|
||||
|
||||
A destination shared between hosts or users, each with its own local index,
|
||||
relies on `prune_grace_period` to keep one's prune from deleting the blobs of
|
||||
@@ -726,6 +732,8 @@ format does and does not protect.
|
||||
5. Delete each directory under `metadata/` with no manifest whose objects are
|
||||
all older than `prune_grace_period`, unless its snapshot is in the local
|
||||
index
|
||||
6. Delete each `.partial` object under `metadata/` and `blobs/` older than
|
||||
`prune_grace_period`
|
||||
|
||||
### chunking and deduplication
|
||||
|
||||
@@ -760,7 +768,7 @@ annotated example also lives in
|
||||
[`configs/config.example.yml`](configs/config.example.yml). Key fields:
|
||||
|
||||
| Field | Default | Description |
|
||||
| -------------------- | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| -------------------- | ------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| `age_recipients` | (required by `snapshot create`) | Age public keys for encryption. Other commands run without one, so a machine that only restores can leave it empty |
|
||||
| `age_secret_key` | (unset) | Age private key for decryption (`snapshot restore`, `snapshot files`, `snapshot verify --deep`). Setting it in the config file places the private key on the backed-up host, defeating the public-key-only design (see "why" above). Prefer the `VAULTIK_AGE_SECRET_KEY` environment variable, supplied only on the machine you restore from. |
|
||||
| `snapshots` | (required) | Named snapshot definitions with paths and excludes |
|
||||
@@ -770,7 +778,7 @@ annotated example also lives in
|
||||
| `chunk_size` | `10MB` | Average chunk size for content-defined chunking |
|
||||
| `blob_size_limit` | `10GB` | Maximum blob size before splitting. Must be at least four times `chunk_size` (the largest chunk the chunker can emit), otherwise a single-chunk blob could exceed the limit. A backup needs free temporary space, because each blob is written whole to a temporary file before it is uploaded (up to about `blob_size_limit`; an rclone destination that cannot stream uploads needs about twice that) and the metadata export writes copies of the local index. Temporary files go to `$TMPDIR` (default `/tmp`); with `TMPDIR` unset, SQLite writes one of those copies to `/var/tmp` |
|
||||
| `compression_level` | `3` | zstd compression level (1-19) |
|
||||
| `prune_grace_period` | `24h` | How old a blob that no snapshot lists must be before `prune` deletes it, and how old every object in a directory under `metadata/` with no manifest must be before `prune` deletes that directory. A backup uploads its blobs and its `db.zst.age` before its manifest, so on a destination shared between hosts or users this must be longer than the longest backup run (see [locking](#locking)). It does not protect a blob a backup reuses from an earlier run that stopped before its manifest: that blob keeps the time of its first upload, so a prune can delete it. A duration such as `36h` or `90m`; `0s` deletes such blobs and directories at once |
|
||||
| `prune_grace_period` | `24h` | How old a blob that no snapshot lists must be before `prune` deletes it, and how old every object in a directory under `metadata/` with no manifest must be before `prune` deletes that directory. `prune` also deletes a `.partial` object left by a killed upload only once it is this old. A backup uploads its blobs and its `db.zst.age` before its manifest, so on a destination shared between hosts or users this must be longer than the longest backup run (see [locking](#locking)). It does not protect a blob a backup reuses from an earlier run that stopped before its manifest: that blob keeps the time of its first upload, so a prune can delete it. A duration such as `36h` or `90m`; `0s` deletes such blobs and directories at once |
|
||||
| `hostname` | system hostname | Hostname used in snapshot IDs |
|
||||
| `index_path` | platform data dir | Local SQLite index path |
|
||||
|
||||
|
||||
@@ -333,6 +333,8 @@ compression_level: 5
|
||||
# How old a blob that no snapshot lists must be before prune deletes it, and
|
||||
# how old every object in a directory under metadata/ with no manifest must
|
||||
# be before prune deletes that directory
|
||||
# prune also deletes a .partial object left by a killed upload only once it is
|
||||
# this old
|
||||
# A backup uploads its blobs and its db.zst.age before its manifest, so on a
|
||||
# destination shared between hosts or users this must be longer than the
|
||||
# longest backup run
|
||||
|
||||
@@ -314,6 +314,9 @@ The restore process doesn't use the local database. Instead:
|
||||
- Delete each directory under `metadata/` that has no `manifest.json.zst`
|
||||
and no local snapshot record, once every object in it is older than
|
||||
`prune_grace_period`
|
||||
3. **Delete Partial Uploads**
|
||||
- Delete each `.partial` object under `metadata/` or `blobs/` left by a
|
||||
killed upload, once it is older than `prune_grace_period`
|
||||
|
||||
### 6. Incomplete Snapshot Cleanup
|
||||
|
||||
|
||||
@@ -329,7 +329,8 @@ The prune operation is safe because:
|
||||
1. It keeps every blob listed in any snapshot's manifest and deletes only blobs
|
||||
that no manifest references. Under `metadata/` it deletes only a directory
|
||||
that has no `manifest.json.zst`, whose name is a remote key, and that no
|
||||
local snapshot record claims
|
||||
local snapshot record claims. It also deletes the `.partial` objects left by
|
||||
killed uploads, which no backup or restore reads
|
||||
2. Manifests are unencrypted and can be read without keys
|
||||
3. If any manifest cannot be downloaded or decoded, prune deletes nothing and
|
||||
exits with an error, rather than treating that snapshot's blobs as
|
||||
@@ -343,8 +344,9 @@ The prune operation is safe because:
|
||||
a backup reuses from an earlier run that stopped before its manifest: the
|
||||
backup skips the upload of a blob already on the destination, so that blob
|
||||
keeps the time of its first upload, and a prune before the backup's manifest
|
||||
is written can delete it; the new snapshot then cannot be restored.
|
||||
`remote nuke` has no such check
|
||||
is written can delete it; the new snapshot then cannot be restored. Prune
|
||||
deletes a `.partial` object only once it is older than `prune_grace_period`,
|
||||
so an upload still running keeps its object. `remote nuke` has no such check
|
||||
|
||||
## Restoration Requirements
|
||||
|
||||
|
||||
@@ -269,9 +269,11 @@ storage_url: ""
|
||||
|
||||
# How old a blob that no snapshot lists must be before prune deletes it,
|
||||
# and how old every object in a directory under metadata/ with no manifest
|
||||
# must be before prune deletes that directory. A backup uploads its blobs
|
||||
# and its db.zst.age before its manifest, so on a destination shared
|
||||
# between hosts or users this must be longer than the longest backup run.
|
||||
# must be before prune deletes that directory. prune also deletes a
|
||||
# .partial object left by a killed upload only once it is this old. A
|
||||
# backup uploads its blobs and its db.zst.age before its manifest, so on
|
||||
# a destination shared between hosts or users this must be longer than
|
||||
# the longest backup run.
|
||||
# It does not protect a blob a backup reuses from an earlier run that
|
||||
# stopped before its manifest: that blob keeps the time of its first
|
||||
# upload, so a prune can delete it.
|
||||
|
||||
@@ -13,9 +13,10 @@ func NewPruneCommand() *cobra.Command {
|
||||
cmd := &cobra.Command{
|
||||
Use: "prune",
|
||||
Short: "Tidy local database and remote storage",
|
||||
Long: `Removes orphaned data from the local index database, and
|
||||
unreferenced blobs and directories under metadata/ with no manifest
|
||||
from the backup destination store.
|
||||
Long: `Removes orphaned data from the local index database. From the
|
||||
backup destination store it removes unreferenced blobs, directories
|
||||
under metadata/ with no manifest, and .partial objects left by killed
|
||||
uploads.
|
||||
|
||||
Local cleanup drops incomplete snapshots and any files, chunks, or
|
||||
blobs no longer referenced by a completed snapshot. Remote cleanup
|
||||
@@ -30,7 +31,8 @@ blob keeps the time of its first upload, so it can be deleted.
|
||||
Remote cleanup also deletes each directory under metadata/ with no
|
||||
manifest, as a backup that stopped before its manifest leaves, once
|
||||
all its objects are older than prune_grace_period and its snapshot is
|
||||
not in the local index.
|
||||
not in the local index. It also deletes each .partial object under
|
||||
metadata/ or blobs/ once it is older than prune_grace_period.
|
||||
|
||||
Snapshot create --prune runs the same cleanup automatically; this
|
||||
command is the manual entry point for the same work (e.g. after a
|
||||
|
||||
@@ -75,7 +75,8 @@ func newRemoteInfoCommand() *cobra.Command {
|
||||
- Size of all snapshot metadata (per snapshot and total)
|
||||
- Count and total size of all blobs
|
||||
- Count and size of referenced blobs (from all manifests)
|
||||
- Count and size of orphaned blobs (not referenced by any manifest)`,
|
||||
- Count and size of orphaned blobs (not referenced by any manifest)
|
||||
- Count and size of .partial objects left by killed uploads`,
|
||||
Args: cobra.NoArgs,
|
||||
RunE: func(cmd *cobra.Command, _ []string) error {
|
||||
// Use unified config resolution
|
||||
|
||||
@@ -138,13 +138,11 @@ on Linux), then /etc/vaultik/config.yml.`,
|
||||
cmd.Flags().BoolVar(&opts.Cron, "cron", false,
|
||||
"Run in cron mode (silent unless warning or error)")
|
||||
cmd.Flags().BoolVar(&opts.Prune, "prune", false,
|
||||
"After backup, drop older snapshots of the same name, remove "+
|
||||
"orphaned blobs older than prune_grace_period, and delete each "+
|
||||
"directory under metadata/ with no manifest once all its "+
|
||||
"objects are older than prune_grace_period and its snapshot is "+
|
||||
"not in the local index (not after a backup that skipped "+
|
||||
"unreadable files, found a configured path empty, or did not "+
|
||||
"find a mount point an earlier snapshot found)")
|
||||
"After backup, drop older snapshots of the same name, then delete "+
|
||||
"from the destination store what prune deletes there (not "+
|
||||
"after a backup that skipped unreadable files, found a "+
|
||||
"configured path empty, or did not find a mount point an "+
|
||||
"earlier snapshot found)")
|
||||
cmd.Flags().StringVar(&opts.KeepNewerThan, "keep-newer-than", "",
|
||||
"With --prune: keep snapshots newer than this duration "+
|
||||
"(e.g. 30d, 4w, 6mo, 1y; m is minutes, mo is months) "+
|
||||
|
||||
@@ -206,7 +206,8 @@ type Config struct {
|
||||
// directory. A backup writes its manifest only once it has uploaded
|
||||
// all its blobs and its db.zst.age, so a younger blob or directory may
|
||||
// belong to a backup still running, on this host or another sharing
|
||||
// the destination store.
|
||||
// the destination store. A prune also deletes a `.partial` object left
|
||||
// by a killed upload only once it is this old.
|
||||
PruneGracePeriod time.Duration `yaml:"prune_grace_period"`
|
||||
|
||||
// StorageURL specifies the storage backend using a URL format.
|
||||
|
||||
@@ -200,9 +200,11 @@ func (f *Storer) ListStream(
|
||||
return f.inner.ListStream(ctx, prefix)
|
||||
}
|
||||
|
||||
// DeletePartialUploads delegates unchanged.
|
||||
func (f *Storer) DeletePartialUploads(ctx context.Context, prefix string) error {
|
||||
return f.inner.DeletePartialUploads(ctx, prefix)
|
||||
// ListPartialUploads delegates unchanged.
|
||||
func (f *Storer) ListPartialUploads(
|
||||
ctx context.Context, prefix string,
|
||||
) ([]storage.ObjectInfo, error) {
|
||||
return f.inner.ListPartialUploads(ctx, prefix)
|
||||
}
|
||||
|
||||
// Info delegates unchanged.
|
||||
|
||||
+26
-13
@@ -56,8 +56,8 @@ const storageDirPerm = 0o755
|
||||
// truncated object at the key a later run would Stat and trust as a complete
|
||||
// blob. The rclone backend's upload does the same on remotes with a
|
||||
// server-side move. List and ListStream skip these files, so a leftover from
|
||||
// an interrupted write is never listed or trusted as a blob; it is otherwise
|
||||
// harmless.
|
||||
// an interrupted write is never listed or trusted as a blob; prune deletes
|
||||
// it once it is older than the prune grace period.
|
||||
const tempSuffix = ".partial"
|
||||
|
||||
// Put stores data at the specified key.
|
||||
@@ -248,22 +248,24 @@ func (f *FileStorer) ListStream(ctx context.Context, prefix string) <-chan Objec
|
||||
return ch
|
||||
}
|
||||
|
||||
// DeletePartialUploads removes every file under prefix whose name ends in
|
||||
// ListPartialUploads returns each file under prefix whose name ends in
|
||||
// tempSuffix. Like List, it reads through symlinked directories and fails
|
||||
// on one whose target is missing. A missing prefix has none to remove.
|
||||
func (f *FileStorer) DeletePartialUploads(ctx context.Context, prefix string) error {
|
||||
basePath := f.fullPath(prefix)
|
||||
|
||||
// on one whose target is missing. A missing prefix has none.
|
||||
func (f *FileStorer) ListPartialUploads(
|
||||
ctx context.Context, prefix string,
|
||||
) ([]ObjectInfo, error) {
|
||||
exists, err := f.prefixExists(prefix)
|
||||
if err != nil {
|
||||
return fmt.Errorf("checking path: %w", err)
|
||||
return nil, fmt.Errorf("checking path: %w", err)
|
||||
}
|
||||
|
||||
if !exists {
|
||||
return nil
|
||||
return nil, nil
|
||||
}
|
||||
|
||||
err = f.walk(basePath, func(path string, info os.FileInfo, err error) error {
|
||||
var partial []ObjectInfo
|
||||
|
||||
err = f.walk(f.fullPath(prefix), func(path string, info os.FileInfo, err error) error {
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
@@ -276,13 +278,24 @@ func (f *FileStorer) DeletePartialUploads(ctx context.Context, prefix string) er
|
||||
return nil
|
||||
}
|
||||
|
||||
return f.fs.Remove(path)
|
||||
})
|
||||
relPath, err := filepath.Rel(f.basePath, path)
|
||||
if err != nil {
|
||||
return fmt.Errorf("walking directory: %w", err)
|
||||
return fmt.Errorf("computing relative path: %w", err)
|
||||
}
|
||||
|
||||
partial = append(partial, ObjectInfo{
|
||||
Key: strings.ReplaceAll(relPath, string(filepath.Separator), "/"),
|
||||
Size: info.Size(),
|
||||
ModTime: info.ModTime(),
|
||||
})
|
||||
|
||||
return nil
|
||||
})
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("walking directory: %w", err)
|
||||
}
|
||||
|
||||
return partial, nil
|
||||
}
|
||||
|
||||
// Info returns human-readable storage location information.
|
||||
|
||||
@@ -121,9 +121,10 @@ func TestFileStorer_ListSkipsPartialFiles(t *testing.T) {
|
||||
}
|
||||
}
|
||||
|
||||
// TestFileStorer_DeletePartialUploads checks that a leftover temp file is
|
||||
// removed and the object at the real key is kept.
|
||||
func TestFileStorer_DeletePartialUploads(t *testing.T) {
|
||||
// TestFileStorer_ListPartialUploads checks that a leftover temp file is
|
||||
// listed with its size and time, that the object at the real key is not,
|
||||
// and that Delete removes the leftover by the listed key.
|
||||
func TestFileStorer_ListPartialUploads(t *testing.T) {
|
||||
t.Parallel()
|
||||
|
||||
base := t.TempDir()
|
||||
@@ -133,23 +134,46 @@ func TestFileStorer_DeletePartialUploads(t *testing.T) {
|
||||
t.Fatalf("NewFileStorer: %v", err)
|
||||
}
|
||||
|
||||
checkListPartialUploads(t, f, base)
|
||||
}
|
||||
|
||||
// checkListPartialUploads stores an object and writes a leftover temp
|
||||
// file next to it in the directory base that s stores into. It checks
|
||||
// that ListPartialUploads lists only the leftover, with its size and a
|
||||
// time, and that Delete removes it by the listed key and keeps the
|
||||
// object.
|
||||
func checkListPartialUploads(t *testing.T, s storage.Storer, base string) {
|
||||
t.Helper()
|
||||
|
||||
ctx := context.Background()
|
||||
|
||||
err = f.Put(ctx, testBlobKey, strings.NewReader("blob-bytes"))
|
||||
err := s.Put(ctx, testBlobKey, strings.NewReader("blob-bytes"))
|
||||
if err != nil {
|
||||
t.Fatalf("Put: %v", err)
|
||||
}
|
||||
|
||||
leftover := filepath.Join(base, testBlobKey+"-123456.partial")
|
||||
leftoverKey := testBlobKey + "-123456.partial"
|
||||
leftover := filepath.Join(base, leftoverKey)
|
||||
|
||||
err = os.WriteFile(leftover, []byte("half"), 0o600)
|
||||
if err != nil {
|
||||
t.Fatalf("writing leftover temp file: %v", err)
|
||||
}
|
||||
|
||||
err = f.DeletePartialUploads(ctx, "blobs/")
|
||||
partial, err := s.ListPartialUploads(ctx, "blobs/")
|
||||
if err != nil {
|
||||
t.Fatalf("DeletePartialUploads: %v", err)
|
||||
t.Fatalf("ListPartialUploads: %v", err)
|
||||
}
|
||||
|
||||
if len(partial) != 1 || partial[0].Key != leftoverKey ||
|
||||
partial[0].Size != int64(len("half")) || partial[0].ModTime.IsZero() {
|
||||
t.Fatalf("ListPartialUploads = %+v, want only %s, 4 bytes, "+
|
||||
"with a time", partial, leftoverKey)
|
||||
}
|
||||
|
||||
err = s.Delete(ctx, partial[0].Key)
|
||||
if err != nil {
|
||||
t.Fatalf("Delete: %v", err)
|
||||
}
|
||||
|
||||
_, err = os.Stat(leftover)
|
||||
@@ -157,7 +181,7 @@ func TestFileStorer_DeletePartialUploads(t *testing.T) {
|
||||
t.Errorf("leftover temp file was not removed: %v", err)
|
||||
}
|
||||
|
||||
_, err = f.Stat(ctx, testBlobKey)
|
||||
_, err = s.Stat(ctx, testBlobKey)
|
||||
if err != nil {
|
||||
t.Errorf("Stat of the real key: %v", err)
|
||||
}
|
||||
|
||||
@@ -239,11 +239,11 @@ func TestFileStorerListFailsOnSymlinkLoop(t *testing.T) {
|
||||
}
|
||||
}
|
||||
|
||||
// TestFileStorerDeletePartialUploadsThroughSymlinkedDirectory checks that
|
||||
// DeletePartialUploads, which remote nuke uses, removes a leftover temp
|
||||
// file behind a symlinked metadata/ or blobs/ directory and keeps the
|
||||
// object at the real key.
|
||||
func TestFileStorerDeletePartialUploadsThroughSymlinkedDirectory(t *testing.T) {
|
||||
// TestFileStorerListPartialUploadsThroughSymlinkedDirectory checks that
|
||||
// ListPartialUploads, which prune and remote nuke use, lists a leftover
|
||||
// temp file behind a symlinked metadata/ or blobs/ directory under its
|
||||
// key, and leaves out the object at the real key.
|
||||
func TestFileStorerListPartialUploadsThroughSymlinkedDirectory(t *testing.T) {
|
||||
t.Parallel()
|
||||
|
||||
ctx := context.Background()
|
||||
@@ -265,29 +265,24 @@ func TestFileStorerDeletePartialUploadsThroughSymlinkedDirectory(t *testing.T) {
|
||||
t.Fatalf("Put: %v", err)
|
||||
}
|
||||
|
||||
leftover := filepath.Join(base, key+"-123456.partial")
|
||||
leftoverKey := key + "-123456.partial"
|
||||
|
||||
err = os.WriteFile(leftover, []byte("half"), 0o600)
|
||||
err = os.WriteFile(filepath.Join(base, leftoverKey),
|
||||
[]byte("half"), 0o600)
|
||||
if err != nil {
|
||||
t.Fatalf("writing leftover temp file: %v", err)
|
||||
}
|
||||
|
||||
moveBehindSymlink(t, base, dir)
|
||||
|
||||
err = s.DeletePartialUploads(ctx, dir+"/")
|
||||
partial, err := s.ListPartialUploads(ctx, dir+"/")
|
||||
if err != nil {
|
||||
t.Fatalf("DeletePartialUploads: %v", err)
|
||||
t.Fatalf("ListPartialUploads: %v", err)
|
||||
}
|
||||
|
||||
// leftover runs through the link.
|
||||
_, err = os.Stat(leftover)
|
||||
if !os.IsNotExist(err) {
|
||||
t.Errorf("leftover temp file was not removed: %v", err)
|
||||
}
|
||||
|
||||
_, err = s.Stat(ctx, key)
|
||||
if err != nil {
|
||||
t.Errorf("Stat of the real key: %v", err)
|
||||
if len(partial) != 1 || partial[0].Key != leftoverKey {
|
||||
t.Errorf("ListPartialUploads = %+v, want only %s",
|
||||
partial, leftoverKey)
|
||||
}
|
||||
})
|
||||
}
|
||||
|
||||
+13
-14
@@ -256,29 +256,28 @@ func (r *RcloneStorer) ListStream(
|
||||
return ch
|
||||
}
|
||||
|
||||
// DeletePartialUploads removes every object under prefix whose name ends
|
||||
// in tempSuffix.
|
||||
func (r *RcloneStorer) DeletePartialUploads(ctx context.Context, prefix string) error {
|
||||
var partial []fs.Object
|
||||
// ListPartialUploads returns each object under prefix whose name ends in
|
||||
// tempSuffix.
|
||||
func (r *RcloneStorer) ListPartialUploads(
|
||||
ctx context.Context, prefix string,
|
||||
) ([]ObjectInfo, error) {
|
||||
var partial []ObjectInfo
|
||||
|
||||
err := operations.ListFn(ctx, r.fsys, func(obj fs.Object) {
|
||||
key := obj.Remote()
|
||||
if strings.HasPrefix(key, prefix) && strings.HasSuffix(key, tempSuffix) {
|
||||
partial = append(partial, obj)
|
||||
partial = append(partial, ObjectInfo{
|
||||
Key: key,
|
||||
Size: obj.Size(),
|
||||
ModTime: obj.ModTime(ctx),
|
||||
})
|
||||
}
|
||||
})
|
||||
if err != nil {
|
||||
return fmt.Errorf("listing objects: %w", err)
|
||||
return nil, fmt.Errorf("listing objects: %w", err)
|
||||
}
|
||||
|
||||
for _, obj := range partial {
|
||||
err = obj.Remove(ctx)
|
||||
if err != nil {
|
||||
return fmt.Errorf("removing object: %w", err)
|
||||
}
|
||||
}
|
||||
|
||||
return nil
|
||||
return partial, nil
|
||||
}
|
||||
|
||||
// Info returns human-readable storage location information.
|
||||
|
||||
@@ -205,45 +205,21 @@ func TestRcloneStorerListSkipsPartialFiles(t *testing.T) {
|
||||
}
|
||||
}
|
||||
|
||||
// TestRcloneStorerDeletePartialUploads checks that a temporary file left
|
||||
// by a killed upload is removed and the object at the real key is kept.
|
||||
// TestRcloneStorerListPartialUploads checks that a temporary file left by
|
||||
// a killed upload is listed with its size and time, that the object at
|
||||
// the real key is not, and that Delete removes the leftover by the listed
|
||||
// key.
|
||||
//
|
||||
//nolint:paralleltest // NewRcloneStorer installs the process-global rclone config
|
||||
func TestRcloneStorerDeletePartialUploads(t *testing.T) {
|
||||
func TestRcloneStorerListPartialUploads(t *testing.T) {
|
||||
dir := t.TempDir()
|
||||
ctx := context.Background()
|
||||
|
||||
s, err := storage.NewRcloneStorer(ctx, ":local", dir)
|
||||
s, err := storage.NewRcloneStorer(context.Background(), ":local", dir)
|
||||
if err != nil {
|
||||
t.Fatalf("NewRcloneStorer: %v", err)
|
||||
}
|
||||
|
||||
err = s.Put(ctx, testBlobKey, strings.NewReader("blob-bytes"))
|
||||
if err != nil {
|
||||
t.Fatalf("Put: %v", err)
|
||||
}
|
||||
|
||||
leftover := filepath.Join(dir, testBlobKey+"-123456.partial")
|
||||
|
||||
err = os.WriteFile(leftover, []byte("half"), 0o600)
|
||||
if err != nil {
|
||||
t.Fatalf("writing leftover temp file: %v", err)
|
||||
}
|
||||
|
||||
err = s.DeletePartialUploads(ctx, "blobs/")
|
||||
if err != nil {
|
||||
t.Fatalf("DeletePartialUploads: %v", err)
|
||||
}
|
||||
|
||||
_, err = os.Stat(leftover)
|
||||
if !os.IsNotExist(err) {
|
||||
t.Errorf("leftover temp file was not removed: %v", err)
|
||||
}
|
||||
|
||||
_, err = s.Stat(ctx, testBlobKey)
|
||||
if err != nil {
|
||||
t.Errorf("Stat of the real key: %v", err)
|
||||
}
|
||||
checkListPartialUploads(t, s, dir)
|
||||
}
|
||||
|
||||
// newRcloneStorerOnWrappedLocal registers name as rclone's local backend
|
||||
|
||||
@@ -100,10 +100,12 @@ func (s *S3Storer) ListStream(ctx context.Context, prefix string) <-chan ObjectI
|
||||
return ch
|
||||
}
|
||||
|
||||
// DeletePartialUploads has nothing to remove: S3 shows an object only once
|
||||
// its upload has completed, so an upload cut off part-way leaves none.
|
||||
func (s *S3Storer) DeletePartialUploads(_ context.Context, _ string) error {
|
||||
return nil
|
||||
// ListPartialUploads finds none: S3 shows an object only once its upload
|
||||
// has completed, so an upload cut off part-way leaves none.
|
||||
func (s *S3Storer) ListPartialUploads(
|
||||
_ context.Context, _ string,
|
||||
) ([]ObjectInfo, error) {
|
||||
return nil, nil
|
||||
}
|
||||
|
||||
// Info returns human-readable storage location information.
|
||||
|
||||
@@ -77,11 +77,12 @@ type Storer interface {
|
||||
// If an error occurs during listing, the final item will have Err set.
|
||||
ListStream(ctx context.Context, prefix string) <-chan ObjectInfo
|
||||
|
||||
// DeletePartialUploads removes every object under prefix that an
|
||||
// upload cut off part-way left under a temporary name ending in
|
||||
// `.partial`. The file and rclone backends' List and ListStream skip
|
||||
// such an object.
|
||||
DeletePartialUploads(ctx context.Context, prefix string) error
|
||||
// ListPartialUploads returns each object under prefix whose name ends
|
||||
// in `.partial`, as a killed upload leaves it, with its key, size and
|
||||
// modification time. The file and rclone
|
||||
// backends' List and ListStream skip such an object; Delete removes
|
||||
// it.
|
||||
ListPartialUploads(ctx context.Context, prefix string) ([]ObjectInfo, error)
|
||||
|
||||
// Info returns human-readable storage location information.
|
||||
Info() Info
|
||||
|
||||
@@ -258,8 +258,10 @@ func (s *stubLister) List(_ context.Context, _ string) ([]string, error) {
|
||||
return nil, errStubUnused
|
||||
}
|
||||
|
||||
func (s *stubLister) DeletePartialUploads(_ context.Context, _ string) error {
|
||||
return errStubUnused
|
||||
func (s *stubLister) ListPartialUploads(
|
||||
_ context.Context, _ string,
|
||||
) ([]storage.ObjectInfo, error) {
|
||||
return nil, errStubUnused
|
||||
}
|
||||
|
||||
func (s *stubLister) Info() storage.Info {
|
||||
|
||||
@@ -216,6 +216,11 @@ type RemoteInfoResult struct {
|
||||
OrphanedBlobCount *int `json:"orphaned_blob_count"`
|
||||
OrphanedBlobSize *int64 `json:"orphaned_blob_size"`
|
||||
|
||||
// `.partial` objects under metadata/ and blobs/ left by killed
|
||||
// uploads. The blob figures above leave them out.
|
||||
PartialUploadCount int `json:"partial_upload_count"`
|
||||
PartialUploadSize int64 `json:"partial_upload_size"`
|
||||
|
||||
// Remote key of each snapshot whose manifest could not be read
|
||||
UnreadableManifests []string `json:"unreadable_manifests,omitempty"`
|
||||
|
||||
@@ -279,6 +284,16 @@ func (v *Vaultik) RemoteInfo(jsonOutput bool) error {
|
||||
return err
|
||||
}
|
||||
|
||||
partialUploads, err := v.listPartialUploads()
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
for _, object := range partialUploads {
|
||||
result.PartialUploadCount++
|
||||
result.PartialUploadSize += object.Size
|
||||
}
|
||||
|
||||
log.Info("Remote info complete",
|
||||
"snapshots", result.TotalMetadataCount,
|
||||
"total_blobs", result.TotalBlobCount,
|
||||
@@ -539,6 +554,12 @@ func (v *Vaultik) printRemoteInfoTable(result *RemoteInfoResult) {
|
||||
"", "", ubytes(result.TotalMetadataSize))
|
||||
}
|
||||
|
||||
v.printRemoteBlobStorage(result)
|
||||
}
|
||||
|
||||
// printRemoteBlobStorage renders the blob storage section of the
|
||||
// human-readable remote info output.
|
||||
func (v *Vaultik) printRemoteBlobStorage(result *RemoteInfoResult) {
|
||||
v.stdoutf("\n=== Blob Storage ===\n")
|
||||
v.stdoutf("Total blobs on remote: %s (%s)\n",
|
||||
humanize.Comma(int64(result.TotalBlobCount)),
|
||||
@@ -552,16 +573,29 @@ func (v *Vaultik) printRemoteInfoTable(result *RemoteInfoResult) {
|
||||
"(%d manifest(s) could not be read, "+
|
||||
"%d manifest(s) under a non-conforming name skipped)\n",
|
||||
len(result.UnreadableManifests), result.SkippedManifestCount)
|
||||
|
||||
return
|
||||
}
|
||||
|
||||
} else {
|
||||
v.stdoutf("Orphaned (unreferenced): %s (%s)\n",
|
||||
humanize.Comma(int64(*result.OrphanedBlobCount)),
|
||||
ubytes(*result.OrphanedBlobSize))
|
||||
}
|
||||
|
||||
if *result.OrphanedBlobCount > 0 {
|
||||
v.stdoutf("\nRun 'vaultik prune' to remove orphaned blobs.\n")
|
||||
v.stdoutf("Partial uploads: %s (%s)\n",
|
||||
humanize.Comma(int64(result.PartialUploadCount)),
|
||||
ubytes(result.PartialUploadSize))
|
||||
|
||||
var toRemove []string
|
||||
|
||||
if result.OrphanedBlobCount != nil && *result.OrphanedBlobCount > 0 {
|
||||
toRemove = append(toRemove, "orphaned blobs")
|
||||
}
|
||||
|
||||
if result.PartialUploadCount > 0 {
|
||||
toRemove = append(toRemove, "partial uploads")
|
||||
}
|
||||
|
||||
if len(toRemove) > 0 {
|
||||
v.stdoutf("\nRun 'vaultik prune' to remove %s.\n",
|
||||
strings.Join(toRemove, " and "))
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -155,10 +155,12 @@ func (m *MockStorer) ListStream(
|
||||
return ch
|
||||
}
|
||||
|
||||
// DeletePartialUploads has nothing to remove: Put stores each object
|
||||
// under its key at once.
|
||||
func (m *MockStorer) DeletePartialUploads(_ context.Context, _ string) error {
|
||||
return nil
|
||||
// ListPartialUploads finds none: Put stores each object under its key at
|
||||
// once.
|
||||
func (m *MockStorer) ListPartialUploads(
|
||||
_ context.Context, _ string,
|
||||
) ([]storage.ObjectInfo, error) {
|
||||
return nil, nil
|
||||
}
|
||||
|
||||
func (m *MockStorer) Info() storage.Info {
|
||||
|
||||
+207
-40
@@ -31,6 +31,11 @@ var errBlobDeleteFailed = errors.New("blob(s) could not be deleted")
|
||||
var errMetadataDirDeleteFailed = errors.New(
|
||||
"metadata director(ies) with no manifest could not be deleted")
|
||||
|
||||
// errPartialUploadDeleteFailed reports `.partial` objects prune could not
|
||||
// delete.
|
||||
var errPartialUploadDeleteFailed = errors.New(
|
||||
"partial upload(s) could not be deleted")
|
||||
|
||||
// errObjectsLeft reports objects remote nuke still finds on the backup
|
||||
// destination store after deleting.
|
||||
var errObjectsLeft = errors.New(
|
||||
@@ -138,9 +143,9 @@ func (v *Vaultik) emptyDestinationStore() error {
|
||||
// that fails is logged and the others still run; that object then
|
||||
// shows up in the listing afterwards.
|
||||
func (v *Vaultik) deleteEverythingUnder(prefix string) ([]string, error) {
|
||||
keys, err := v.Storage.List(v.ctx, prefix)
|
||||
keys, err := v.listEverythingUnder(prefix)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("listing %s: %w", prefix, err)
|
||||
return nil, err
|
||||
}
|
||||
|
||||
for _, key := range keys {
|
||||
@@ -150,19 +155,47 @@ func (v *Vaultik) deleteEverythingUnder(prefix string) ([]string, error) {
|
||||
}
|
||||
}
|
||||
|
||||
// The file and rclone listings skip `.partial` objects, so the
|
||||
// deletes above never reach them.
|
||||
err = v.Storage.DeletePartialUploads(v.ctx, prefix)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("deleting partial uploads: %w", err)
|
||||
}
|
||||
return v.listEverythingUnder(prefix)
|
||||
}
|
||||
|
||||
left, err := v.Storage.List(v.ctx, prefix)
|
||||
// listEverythingUnder returns the key of every object under prefix,
|
||||
// including the `.partial` objects left by killed uploads, which the file
|
||||
// and rclone backends' List skips.
|
||||
func (v *Vaultik) listEverythingUnder(prefix string) ([]string, error) {
|
||||
keys, err := v.Storage.List(v.ctx, prefix)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("listing %s: %w", prefix, err)
|
||||
}
|
||||
|
||||
return left, nil
|
||||
partialUploads, err := v.Storage.ListPartialUploads(v.ctx, prefix)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("listing partial uploads under %s: %w",
|
||||
prefix, err)
|
||||
}
|
||||
|
||||
for _, object := range partialUploads {
|
||||
keys = append(keys, object.Key)
|
||||
}
|
||||
|
||||
return keys, nil
|
||||
}
|
||||
|
||||
// listPartialUploads returns the `.partial` objects under metadata/ and
|
||||
// blobs/ left by killed uploads.
|
||||
func (v *Vaultik) listPartialUploads() ([]storage.ObjectInfo, error) {
|
||||
var partialUploads []storage.ObjectInfo
|
||||
|
||||
for _, prefix := range []string{"metadata/", "blobs/"} {
|
||||
objects, err := v.Storage.ListPartialUploads(v.ctx, prefix)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("listing partial uploads under %s: %w",
|
||||
prefix, err)
|
||||
}
|
||||
|
||||
partialUploads = append(partialUploads, objects...)
|
||||
}
|
||||
|
||||
return partialUploads, nil
|
||||
}
|
||||
|
||||
// PruneBlobsResult contains the result of a blob prune operation
|
||||
@@ -182,11 +215,19 @@ type PruneBlobsResult struct {
|
||||
// not delete.
|
||||
MetadataDirsDeleted int `json:"metadata_dirs_deleted"`
|
||||
MetadataDirsFailed int `json:"metadata_dirs_failed,omitempty"`
|
||||
// PartialUploadsDeleted counts the `.partial` objects prune deleted,
|
||||
// and PartialUploadsFailed those it could not delete.
|
||||
// PartialUploadsTooNew counts those kept for being newer than the
|
||||
// prune grace period, or of unknown age.
|
||||
PartialUploadsDeleted int `json:"partial_uploads_deleted"`
|
||||
PartialUploadsFailed int `json:"partial_uploads_failed,omitempty"`
|
||||
PartialUploadsTooNew int `json:"partial_uploads_too_new"`
|
||||
}
|
||||
|
||||
// Prune removes orphaned data from the local index database AND, from the
|
||||
// backup destination store, unreferenced blobs and the directories under
|
||||
// metadata/ with no manifest. This is the single user-facing prune entry
|
||||
// backup destination store, unreferenced blobs, the directories under
|
||||
// metadata/ with no manifest and the `.partial` objects left by killed
|
||||
// uploads. This is the single user-facing prune entry
|
||||
// point — the split between local and remote cleanup is an
|
||||
// implementation detail. Calling code should prefer this method over
|
||||
// PruneDatabase or PruneBlobs individually unless it specifically wants
|
||||
@@ -214,8 +255,9 @@ func (v *Vaultik) Prune(opts *PruneOptions) error {
|
||||
return v.PruneBlobs(opts)
|
||||
}
|
||||
|
||||
// PruneBlobs removes unreferenced blobs from storage, and each directory
|
||||
// under metadata/ that a backup left with no manifest.
|
||||
// PruneBlobs removes unreferenced blobs from storage, each directory
|
||||
// under metadata/ that a backup left with no manifest, and each
|
||||
// `.partial` object left by a killed upload.
|
||||
func (v *Vaultik) PruneBlobs(opts *PruneOptions) error {
|
||||
log.Info("Starting prune operation")
|
||||
|
||||
@@ -242,14 +284,21 @@ func (v *Vaultik) PruneBlobs(opts *PruneOptions) error {
|
||||
return err
|
||||
}
|
||||
|
||||
partialUploads, partialTooNew, err := v.findOldPartialUploads()
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
result := &PruneBlobsResult{
|
||||
BlobsFound: len(unreferencedBlobs),
|
||||
BlobsTooNew: tooNew,
|
||||
PartialUploadsTooNew: partialTooNew,
|
||||
}
|
||||
|
||||
v.reportBlobsTooNew(tooNew, opts)
|
||||
v.reportTooNew(result, opts)
|
||||
|
||||
if len(unreferencedBlobs) == 0 && len(dirsWithoutManifest) == 0 {
|
||||
if len(unreferencedBlobs) == 0 && len(dirsWithoutManifest) == 0 &&
|
||||
len(partialUploads) == 0 {
|
||||
log.Info("No unreferenced blobs found")
|
||||
|
||||
if opts.JSON {
|
||||
@@ -263,15 +312,17 @@ func (v *Vaultik) PruneBlobs(opts *PruneOptions) error {
|
||||
|
||||
log.Info("Found unreferenced blobs",
|
||||
"count", len(unreferencedBlobs), "total_size", ubytes(totalSize),
|
||||
"metadata_dirs_without_manifest", len(dirsWithoutManifest))
|
||||
"metadata_dirs_without_manifest", len(dirsWithoutManifest),
|
||||
"partial_uploads", len(partialUploads))
|
||||
|
||||
if !v.confirmPrune(opts, len(unreferencedBlobs), totalSize,
|
||||
dirsWithoutManifest) {
|
||||
dirsWithoutManifest, partialUploads) {
|
||||
return ErrNotConfirmed
|
||||
}
|
||||
|
||||
v.deleteUnreferencedBlobs(unreferencedBlobs, allBlobs, result)
|
||||
v.deleteMetadataDirs(dirsWithoutManifest, result)
|
||||
v.deletePartialUploads(partialUploads, result)
|
||||
|
||||
if opts.JSON {
|
||||
err = v.outputPruneBlobsJSON(result)
|
||||
@@ -279,9 +330,16 @@ func (v *Vaultik) PruneBlobs(opts *PruneOptions) error {
|
||||
return err
|
||||
}
|
||||
} else {
|
||||
v.reportPruneDeletions(result, len(dirsWithoutManifest))
|
||||
v.reportPruneDeletions(result, len(dirsWithoutManifest),
|
||||
len(partialUploads))
|
||||
}
|
||||
|
||||
return pruneDeleteFailures(result)
|
||||
}
|
||||
|
||||
// pruneDeleteFailures returns an error for each kind of object prune
|
||||
// could not delete, or nil when every delete succeeded.
|
||||
func pruneDeleteFailures(result *PruneBlobsResult) error {
|
||||
var failures []error
|
||||
|
||||
if result.BlobsFailed > 0 {
|
||||
@@ -294,26 +352,41 @@ func (v *Vaultik) PruneBlobs(opts *PruneOptions) error {
|
||||
result.MetadataDirsFailed, errMetadataDirDeleteFailed))
|
||||
}
|
||||
|
||||
if result.PartialUploadsFailed > 0 {
|
||||
failures = append(failures, fmt.Errorf("%d %w",
|
||||
result.PartialUploadsFailed, errPartialUploadDeleteFailed))
|
||||
}
|
||||
|
||||
return errors.Join(failures...)
|
||||
}
|
||||
|
||||
// confirmPrune states what prune found to delete, unreferenced blobs and
|
||||
// confirmPrune states what prune found to delete, unreferenced blobs,
|
||||
// directories under metadata/ with no manifest, naming each directory by
|
||||
// its remote key, and, unless --force is given, asks the operator to
|
||||
// confirm. Under --json it prints nothing and does not ask. It reports
|
||||
// whether to go ahead.
|
||||
// its remote key, and `.partial` objects, and, unless --force is given,
|
||||
// asks the operator to confirm. Under --json it prints nothing and does
|
||||
// not ask. It reports whether to go ahead.
|
||||
func (v *Vaultik) confirmPrune(
|
||||
opts *PruneOptions, blobCount int, blobsSize int64, dirs []string,
|
||||
partialUploads []storage.ObjectInfo,
|
||||
) bool {
|
||||
if opts.JSON {
|
||||
return true
|
||||
}
|
||||
|
||||
var partialSize int64
|
||||
|
||||
for _, object := range partialUploads {
|
||||
partialSize += object.Size
|
||||
}
|
||||
|
||||
dirCount := len(dirs)
|
||||
partialCount := len(partialUploads)
|
||||
blobsFound := fmt.Sprintf("Found %d unreferenced blob(s) totaling %s.",
|
||||
blobCount, ubytes(blobsSize))
|
||||
dirsFound := fmt.Sprintf("Found %d metadata director(ies) left by "+
|
||||
"backups that stopped before their manifest:", dirCount)
|
||||
partialFound := fmt.Sprintf("Found %d partial upload(s) totaling %s, "+
|
||||
"left by uploads cut off part-way.", partialCount, ubytes(partialSize))
|
||||
|
||||
if opts.Force {
|
||||
if blobCount > 0 {
|
||||
@@ -328,6 +401,10 @@ func (v *Vaultik) confirmPrune(
|
||||
}
|
||||
}
|
||||
|
||||
if partialCount > 0 {
|
||||
v.UI.Infof("%s", partialFound)
|
||||
}
|
||||
|
||||
return true
|
||||
}
|
||||
|
||||
@@ -346,16 +423,13 @@ func (v *Vaultik) confirmPrune(
|
||||
}
|
||||
}
|
||||
|
||||
switch {
|
||||
case dirCount == 0:
|
||||
v.stdoutf("\nDelete %d unreferenced blob(s)? [y/N] ", blobCount)
|
||||
case blobCount == 0:
|
||||
v.stdoutf("\nDelete %d metadata director(ies)? [y/N] ", dirCount)
|
||||
default:
|
||||
v.stdoutf("\nDelete %d unreferenced blob(s) and %d metadata "+
|
||||
"director(ies)? [y/N] ", blobCount, dirCount)
|
||||
if partialCount > 0 {
|
||||
v.stdoutf("%s\n", partialFound)
|
||||
}
|
||||
|
||||
v.stdoutf("\nDelete %s? [y/N] ",
|
||||
pruneTargets(blobCount, dirCount, partialCount))
|
||||
|
||||
var confirm string
|
||||
|
||||
// A failure to read the answer means no confirmation.
|
||||
@@ -369,9 +443,36 @@ func (v *Vaultik) confirmPrune(
|
||||
return true
|
||||
}
|
||||
|
||||
// reportPruneDeletions says how many blobs and metadata directories
|
||||
// prune deleted, naming each kind only when it found some to delete.
|
||||
func (v *Vaultik) reportPruneDeletions(result *PruneBlobsResult, dirCount int) {
|
||||
// pruneTargets names what the prune prompt offers to delete, such as
|
||||
// "2 unreferenced blob(s) and 1 partial upload(s)", leaving out each kind
|
||||
// with a count of zero.
|
||||
func pruneTargets(blobCount, dirCount, partialCount int) string {
|
||||
var targets []string
|
||||
|
||||
if blobCount > 0 {
|
||||
targets = append(targets,
|
||||
fmt.Sprintf("%d unreferenced blob(s)", blobCount))
|
||||
}
|
||||
|
||||
if dirCount > 0 {
|
||||
targets = append(targets,
|
||||
fmt.Sprintf("%d metadata director(ies)", dirCount))
|
||||
}
|
||||
|
||||
if partialCount > 0 {
|
||||
targets = append(targets,
|
||||
fmt.Sprintf("%d partial upload(s)", partialCount))
|
||||
}
|
||||
|
||||
return strings.Join(targets, " and ")
|
||||
}
|
||||
|
||||
// reportPruneDeletions says how many blobs, metadata directories and
|
||||
// partial uploads prune deleted, naming each kind only when it found some
|
||||
// to delete.
|
||||
func (v *Vaultik) reportPruneDeletions(
|
||||
result *PruneBlobsResult, dirCount, partialCount int,
|
||||
) {
|
||||
if result.BlobsFound > 0 {
|
||||
v.UI.Infof("Deleted %d blob(s) totaling %s.",
|
||||
result.BlobsDeleted, ubytes(result.BytesFreed))
|
||||
@@ -381,6 +482,11 @@ func (v *Vaultik) reportPruneDeletions(result *PruneBlobsResult, dirCount int) {
|
||||
v.UI.Infof("Deleted %d metadata director(ies) with no manifest.",
|
||||
result.MetadataDirsDeleted)
|
||||
}
|
||||
|
||||
if partialCount > 0 {
|
||||
v.UI.Infof("Deleted %d partial upload(s).",
|
||||
result.PartialUploadsDeleted)
|
||||
}
|
||||
}
|
||||
|
||||
// listMetadataObjects lists every object under metadata/ on the
|
||||
@@ -611,18 +717,79 @@ func (v *Vaultik) deleteMetadataDirs(
|
||||
}
|
||||
}
|
||||
|
||||
// reportBlobsTooNew says how many unreferenced blobs prune kept for being
|
||||
// newer than the grace period or of unknown age. Under --json the
|
||||
// document carries the count instead.
|
||||
func (v *Vaultik) reportBlobsTooNew(tooNew int, opts *PruneOptions) {
|
||||
if tooNew == 0 || opts.JSON {
|
||||
// reportTooNew says how many unreferenced blobs and partial uploads prune
|
||||
// kept for being newer than the grace period or of unknown age. Under
|
||||
// --json the document carries the counts instead.
|
||||
func (v *Vaultik) reportTooNew(result *PruneBlobsResult, opts *PruneOptions) {
|
||||
if opts.JSON {
|
||||
return
|
||||
}
|
||||
|
||||
if result.BlobsTooNew > 0 {
|
||||
v.UI.Infof("Kept %d unreferenced blob(s) newer than "+
|
||||
"prune_grace_period (%s), or of unknown age: a backup still "+
|
||||
"running may need them.",
|
||||
tooNew, v.Config.PruneGracePeriod)
|
||||
result.BlobsTooNew, v.Config.PruneGracePeriod)
|
||||
}
|
||||
|
||||
if result.PartialUploadsTooNew > 0 {
|
||||
v.UI.Infof("Kept %d partial upload(s) newer than "+
|
||||
"prune_grace_period (%s), or of unknown age: an upload may "+
|
||||
"still be writing them.",
|
||||
result.PartialUploadsTooNew, v.Config.PruneGracePeriod)
|
||||
}
|
||||
}
|
||||
|
||||
// findOldPartialUploads returns the `.partial` objects under metadata/
|
||||
// and blobs/ that are older than the prune grace period. It also returns
|
||||
// how many it left out for being newer than that, or of unknown age
|
||||
// because the listing gave no time for them.
|
||||
func (v *Vaultik) findOldPartialUploads() ([]storage.ObjectInfo, int, error) {
|
||||
partialUploads, err := v.listPartialUploads()
|
||||
if err != nil {
|
||||
return nil, 0, err
|
||||
}
|
||||
|
||||
cutoff := time.Now().UTC().Add(-v.Config.PruneGracePeriod)
|
||||
|
||||
var (
|
||||
old []storage.ObjectInfo
|
||||
tooNew int
|
||||
)
|
||||
|
||||
for _, object := range partialUploads {
|
||||
if object.ModTime.IsZero() || object.ModTime.After(cutoff) {
|
||||
tooNew++
|
||||
|
||||
continue
|
||||
}
|
||||
|
||||
old = append(old, object)
|
||||
}
|
||||
|
||||
return old, tooNew, nil
|
||||
}
|
||||
|
||||
// deletePartialUploads deletes the given `.partial` objects and counts
|
||||
// in the result those deleted and those it could not delete.
|
||||
func (v *Vaultik) deletePartialUploads(
|
||||
partialUploads []storage.ObjectInfo, result *PruneBlobsResult,
|
||||
) {
|
||||
for _, object := range partialUploads {
|
||||
err := v.Storage.Delete(v.ctx, object.Key)
|
||||
if err != nil {
|
||||
log.Error("Failed to delete partial upload",
|
||||
"key", object.Key, "error", err)
|
||||
|
||||
result.PartialUploadsFailed++
|
||||
|
||||
continue
|
||||
}
|
||||
|
||||
log.Info("Deleted partial upload", "key", object.Key)
|
||||
|
||||
result.PartialUploadsDeleted++
|
||||
}
|
||||
}
|
||||
|
||||
// deleteUnreferencedBlobs deletes the given blobs from storage and
|
||||
|
||||
@@ -0,0 +1,157 @@
|
||||
package vaultik_test
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"context"
|
||||
"encoding/json"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/stretchr/testify/assert"
|
||||
"github.com/stretchr/testify/require"
|
||||
"sneak.berlin/go/vaultik/internal/config"
|
||||
"sneak.berlin/go/vaultik/internal/log"
|
||||
"sneak.berlin/go/vaultik/internal/storage"
|
||||
"sneak.berlin/go/vaultik/internal/storage/faultstore"
|
||||
"sneak.berlin/go/vaultik/internal/ui"
|
||||
"sneak.berlin/go/vaultik/internal/vaultik"
|
||||
)
|
||||
|
||||
// partialUploadGracePeriod is the prune grace period of the vaultik
|
||||
// newPartialUploadStore returns.
|
||||
const partialUploadGracePeriod = time.Hour
|
||||
|
||||
// newPartialUploadStore returns a vaultik on a file:// destination
|
||||
// holding `.partial` files left by killed uploads: one under blobs/
|
||||
// and one under metadata/ older than the prune grace period, and one
|
||||
// under blobs/ newer than that. It also returns the buffer the vaultik
|
||||
// writes its output to, the paths of the old files and the path of the
|
||||
// new one.
|
||||
func newPartialUploadStore(
|
||||
t *testing.T,
|
||||
) (*vaultik.Vaultik, *bytes.Buffer, []string, string) {
|
||||
t.Helper()
|
||||
|
||||
storeDir := t.TempDir()
|
||||
|
||||
store, err := storage.NewFileStorer(storeDir)
|
||||
require.NoError(t, err)
|
||||
|
||||
hash := testBlobHashA
|
||||
oldFiles := []string{
|
||||
filepath.Join(storeDir, "blobs", hash[:2], hash[2:4],
|
||||
hash+"-123456.partial"),
|
||||
filepath.Join(storeDir, metadataObjectKey(
|
||||
"testhost_home_2026-01-01T00:00:00Z", "db.zst.age-123456.partial")),
|
||||
}
|
||||
newFile := filepath.Join(storeDir, "blobs", hash[:2], hash[2:4],
|
||||
hash+"-654321.partial")
|
||||
|
||||
beforeGracePeriod := time.Now().Add(-partialUploadGracePeriod - time.Minute)
|
||||
|
||||
for _, path := range append([]string{newFile}, oldFiles...) {
|
||||
require.NoError(t, os.MkdirAll(filepath.Dir(path), 0o750))
|
||||
require.NoError(t, os.WriteFile(path, []byte("half an upload"), 0o600))
|
||||
}
|
||||
|
||||
for _, path := range oldFiles {
|
||||
require.NoError(t, os.Chtimes(path, beforeGracePeriod, beforeGracePeriod))
|
||||
}
|
||||
|
||||
stdout := &bytes.Buffer{}
|
||||
v := &vaultik.Vaultik{
|
||||
Config: &config.Config{PruneGracePeriod: partialUploadGracePeriod},
|
||||
Storage: store,
|
||||
Stdout: stdout,
|
||||
UI: ui.NewWithColor(stdout, false),
|
||||
}
|
||||
v.SetContext(context.Background())
|
||||
|
||||
return v, stdout, oldFiles, newFile
|
||||
}
|
||||
|
||||
// TestPruneDeletesPartialUploadsOlderThanGracePeriod checks that prune
|
||||
// deletes a `.partial` file under blobs/ or metadata/ left by a killed
|
||||
// upload once it is older than the grace period, keeps a newer one,
|
||||
// and counts both in the --json document.
|
||||
func TestPruneDeletesPartialUploadsOlderThanGracePeriod(t *testing.T) {
|
||||
log.Initialize(log.Config{})
|
||||
t.Parallel()
|
||||
|
||||
v, stdout, oldFiles, newFile := newPartialUploadStore(t)
|
||||
|
||||
require.NoError(t, v.PruneBlobs(&vaultik.PruneOptions{JSON: true}))
|
||||
|
||||
for _, path := range oldFiles {
|
||||
_, err := os.Stat(path)
|
||||
require.ErrorIs(t, err, os.ErrNotExist, "%s must be deleted", path)
|
||||
}
|
||||
|
||||
_, err := os.Stat(newFile)
|
||||
require.NoError(t, err, "a partial upload newer than the grace period "+
|
||||
"must be kept")
|
||||
|
||||
var result vaultik.PruneBlobsResult
|
||||
|
||||
require.NoError(t, json.Unmarshal(stdout.Bytes(), &result))
|
||||
assert.Equal(t, 2, result.PartialUploadsDeleted)
|
||||
assert.Equal(t, 1, result.PartialUploadsTooNew)
|
||||
}
|
||||
|
||||
// TestPruneFailedPartialUploadDeleteReturnsError checks that prune
|
||||
// returns an error, and counts the failures in the --json document, when
|
||||
// the `.partial` files it found cannot be deleted.
|
||||
func TestPruneFailedPartialUploadDeleteReturnsError(t *testing.T) {
|
||||
log.Initialize(log.Config{})
|
||||
t.Parallel()
|
||||
|
||||
v, stdout, oldFiles, _ := newPartialUploadStore(t)
|
||||
|
||||
store := faultstore.New(v.Storage)
|
||||
store.OnDelete = func(string) bool { return true }
|
||||
v.Storage = store
|
||||
|
||||
err := v.PruneBlobs(&vaultik.PruneOptions{JSON: true})
|
||||
|
||||
require.ErrorContains(t, err, "2 partial upload(s) could not be deleted")
|
||||
|
||||
for _, path := range oldFiles {
|
||||
_, statErr := os.Stat(path)
|
||||
require.NoError(t, statErr, "%s must still exist", path)
|
||||
}
|
||||
|
||||
var result vaultik.PruneBlobsResult
|
||||
|
||||
require.NoError(t, json.Unmarshal(stdout.Bytes(), &result),
|
||||
"stdout must hold the JSON document, got:\n%s", stdout.String())
|
||||
assert.Equal(t, 2, result.PartialUploadsFailed)
|
||||
assert.Equal(t, 0, result.PartialUploadsDeleted)
|
||||
}
|
||||
|
||||
// TestRemoteInfoCountsPartialUploads checks that remote info counts
|
||||
// every `.partial` file under blobs/ and metadata/, whatever its age, and
|
||||
// advises running prune.
|
||||
func TestRemoteInfoCountsPartialUploads(t *testing.T) {
|
||||
log.Initialize(log.Config{})
|
||||
t.Parallel()
|
||||
|
||||
v, stdout, _, _ := newPartialUploadStore(t)
|
||||
|
||||
require.NoError(t, v.RemoteInfo(true))
|
||||
|
||||
var result vaultik.RemoteInfoResult
|
||||
|
||||
require.NoError(t, json.Unmarshal(stdout.Bytes(), &result))
|
||||
assert.Equal(t, 3, result.PartialUploadCount)
|
||||
assert.Equal(t, int64(3*len("half an upload")), result.PartialUploadSize)
|
||||
assert.Equal(t, 0, result.TotalBlobCount)
|
||||
|
||||
stdout.Reset()
|
||||
require.NoError(t, v.RemoteInfo(false))
|
||||
|
||||
text := stdout.String()
|
||||
assert.Contains(t, text, "Partial uploads: 3 (")
|
||||
assert.Contains(t, text, "Run 'vaultik prune' to remove partial uploads.")
|
||||
}
|
||||
@@ -136,10 +136,12 @@ func (s *testStorer) ListStream(
|
||||
return ch
|
||||
}
|
||||
|
||||
// DeletePartialUploads has nothing to remove: Put stores each object
|
||||
// under its key at once.
|
||||
func (s *testStorer) DeletePartialUploads(_ context.Context, _ string) error {
|
||||
return nil
|
||||
// ListPartialUploads finds none: Put stores each object under its key at
|
||||
// once.
|
||||
func (s *testStorer) ListPartialUploads(
|
||||
_ context.Context, _ string,
|
||||
) ([]storage.ObjectInfo, error) {
|
||||
return nil, nil
|
||||
}
|
||||
|
||||
func (s *testStorer) Info() storage.Info {
|
||||
|
||||
@@ -302,10 +302,10 @@ func (v *Vaultik) snapshotNamesToCreate(requested []string) ([]string, error) {
|
||||
}
|
||||
|
||||
// runPostBackupPrune drops older snapshots of the given names and removes
|
||||
// from remote storage the orphan blobs and the directories under
|
||||
// metadata/ with no manifest, as PruneBlobs does. If keepNewerThan is set
|
||||
// (e.g. "4w"), snapshots newer than that duration are kept. Otherwise only
|
||||
// the latest snapshot of each name is kept.
|
||||
// from remote storage the orphan blobs, the directories under metadata/
|
||||
// with no manifest and the `.partial` objects, as PruneBlobs does. If
|
||||
// keepNewerThan is set (e.g. "4w"), snapshots newer than that duration are
|
||||
// kept. Otherwise only the latest snapshot of each name is kept.
|
||||
func (v *Vaultik) runPostBackupPrune(
|
||||
snapshotNames []string, keepNewerThan string,
|
||||
) error {
|
||||
|
||||
Reference in new issue
Block a user