Make prune and remote nuke fail when a delete fails (closes #290 )

prune, and the prune step of snapshot create --prune, counted a blob it could not delete and still exited 0. It now prints its summary or --json document and then returns an error, which replaces the "Failed to delete N blob(s)" line. remote nuke only warned when a snapshot's metadata could not be deleted, advised vaultik prune, which never removes metadata, and then reported the store empty. It now stops with an error, does not report the store empty, and says to run vaultik remote nuke --force again. It checks that the local index belongs to the configured destination before deleting anything, and that error carries no such advice. A successful nuke no longer advises prune. Model: opus-5-5
2026-10-13 05:44:12 +00:00
17 changed files with 319 additions and 362 deletions
@@ -286,9 +286,7 @@ needed.
  remove orphaned blobs from remote storage. By default keeps only the latest
  snapshot per name; use `--keep-newer-than` for a rolling window.
 * `--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 before the backup starts.
+  this duration instead of only the latest (e.g. `4w`, `30d`, `6mo`, `1y`)

 **`snapshot list`**: Show every snapshot known to this host — the union
 of the local index and the backup destination store — with timestamps
@@ -386,10 +384,12 @@ manifest on the destination store to compute the still-referenced blob
 set and delete any blob not in that set. 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.
+manual entry point for the same work. If a blob 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
-  `--force` does.
+  `--force` does. A blob that cannot be deleted is counted in
+  `blobs_failed`, and the document is printed before the non-zero exit.

 **`info`**: Display system configuration, storage settings, encryption
 recipients, and local database statistics.
@@ -403,15 +403,14 @@ orphaned blob figures are reported as unknown; `--json` gives them as
 `null`, lists the remote key of each unreadable manifest 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.
+snapshot's blob count and blob size unknown, `null` in `--json`.
 * `--json`: Output as JSON

 **`remote nuke`**: Delete every snapshot's metadata and every blob from the
 backup destination store, leaving the bucket prefix empty. Destructive and
 irreversible. This is the single supported way to wipe the entire
-destination store.
+destination store. If anything cannot be deleted, the command exits
+non-zero; running it again deletes what is left.
 * `--force`: Required to confirm destruction.

 **`database delete`**: Delete the local SQLite state database file
@@ -22,13 +22,16 @@ the tag exists and is exercised; what is left is merging `next` to

 # Completed Steps

- 2026-10-13: Stopped `remote info` counting a file under `blobs/` that
-  is not a blob as an orphaned blob
-  ([issue #295](https://git.eeqj.de/sneak/vaultik/issues/295)). A stray
-  file such as `.DS_Store` was reported as an orphaned blob and the
-  report advised running `prune`, which skips such a file and deleted
-  nothing. `remote info` now takes its blob list from the same listing
-  `prune` uses, in the table and in `--json`.
+- 2026-10-13: Made `prune` and `remote nuke` exit non-zero when a delete
+  fails ([issue #290](https://git.eeqj.de/sneak/vaultik/issues/290)).
+  `prune`, and the prune step of `snapshot create --prune`, counted a
+  blob it could not delete and still exited 0; it now prints its summary
+  or `--json` document and then fails. `remote nuke` warned about a
+  snapshot's metadata it could not delete, said to run `vaultik prune`,
+  which never removes metadata, and then reported the store empty. It
+  now fails without that report and says to run `vaultik remote nuke
+  --force` again, and a successful nuke no longer advises `prune` for
+  blobs it deletes itself.

 - 2026-10-13: Made `snapshot remove` fail for a snapshot that does not
  exist ([issue #293](https://git.eeqj.de/sneak/vaultik/issues/293)).
@@ -90,7 +90,7 @@ Stores information about packed, compressed, and encrypted blob files.
 - `created_ts` (INTEGER NOT NULL) - Creation timestamp
 - `finished_ts` (INTEGER) - Finalization timestamp (NULL if in progress)
 - `uncompressed_size` (INTEGER NOT NULL DEFAULT 0) - Total size of chunks before compression
- `compressed_size` (INTEGER NOT NULL DEFAULT 0) - Size after compression and encryption. Once uploaded, the size of the object in storage: when the blob was already stored, under another compression level or other recipients, this can differ from the size of the local copy
+- `compressed_size` (INTEGER NOT NULL DEFAULT 0) - Size after compression and encryption
 - `uploaded_ts` (INTEGER) - Upload completion timestamp (NULL if not uploaded)

 **Purpose:** Tracks blob lifecycle from creation through upload. The UUID primary key allows immediate association of chunks with blobs.
@@ -183,7 +183,7 @@ Tracks blob upload metrics.
   - `UPDATE blobs SET blob_hash = ?, finished_ts = ?` - Finalize blob after packing

 4. **Upload**
-   - `UPDATE blobs SET uploaded_ts = ?, compressed_size = ?` - Mark blob as uploaded, with the size of the object in storage (the upload is skipped when the blob is already stored)
+   - `UPDATE blobs SET uploaded_ts = ?` - Mark blob as uploaded
   - `INSERT INTO uploads` - Record upload metrics with snapshot_id
   - `INSERT INTO snapshot_blobs` - Associate blob with snapshot

@@ -1,108 +0,0 @@
-package cli //nolint:testpackage // shares runEntry and the hermetic config helpers
-
-import (
-	"os"
-	"path/filepath"
-	"strings"
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-// Flags of `snapshot create` used to build argument vectors below.
-const (
-	flagPrune         = "--prune"
-	flagKeepNewerThan = "--keep-newer-than"
-)
-
-// TestEntrySnapshotCreateRejectsKeepNewerThanBeforeBackup checks that a
-// --keep-newer-than the post-backup prune could not use stops `snapshot
-// create` before the backup: the command fails and the destination
-// store stays empty. It used to be read only by the prune, after the
-// snapshot was uploaded, and without --prune it was ignored.
-//
-//nolint:paralleltest // replaces os.Args, os.Stdout, os.Stderr and the xdg globals
-func TestEntrySnapshotCreateRejectsKeepNewerThanBeforeBackup(t *testing.T) {
-	for _, testCase := range []struct {
-		name string
-		args []string
-	}{
-		{
-			name: "decimal in a day count",
-			args: []string{flagPrune, flagKeepNewerThan, "1.5d"},
-		},
-		{
-			name: "unknown unit",
-			args: []string{flagPrune, flagKeepNewerThan, "4x"},
-		},
-		{
-			name: "without --prune",
-			args: []string{flagKeepNewerThan, "4w"},
-		},
-	} {
-		t.Run(testCase.name, func(t *testing.T) {
-			configPath, storeDir := writeBackupConfig(t)
-
-			code, _, _ := runEntry(t, append(
-				[]string{flagConfig, configPath, cmdSnapshot, cmdCreate},
-				testCase.args...)...)
-
-			assert.Equal(t, 1, code)
-
-			stored, err := os.ReadDir(storeDir)
-			require.NoError(t, err)
-			assert.Empty(t, stored, "nothing may be uploaded")
-		})
-	}
-}
-
-// TestEntrySnapshotCreateReportsPostBackupPruneFailure makes the
-// post-backup prune fail after a good backup, by planting a manifest
-// that cannot be decoded, and checks that the failure is reported as the
-// prune's and not as a failed snapshot creation.
-//
-//nolint:paralleltest // replaces os.Args, os.Stdout, os.Stderr and the xdg globals
-func TestEntrySnapshotCreateReportsPostBackupPruneFailure(t *testing.T) {
-	configPath, storeDir := writeBackupConfig(t)
-
-	// A remote key is 64 lowercase hex characters. The prune reads every
-	// manifest on the store and fails on one it cannot decode.
-	plantedDir := filepath.Join(storeDir, "metadata", strings.Repeat("a", 64))
-	require.NoError(t, os.MkdirAll(plantedDir, 0o750))
-	require.NoError(t, os.WriteFile(
-		filepath.Join(plantedDir, "manifest.json.zst"),
-		[]byte("not a valid manifest"), 0o600))
-
-	code, _, stderr := runEntry(t,
-		flagConfig, configPath, cmdSnapshot, cmdCreate, flagPrune)
-
-	assert.Equal(t, 1, code)
-	assert.Contains(t, stderr, "Backup completed, but post-backup prune failed")
-	assert.NotContains(t, stderr, "Snapshot creation failed")
-
-	snapshots, err := os.ReadDir(filepath.Join(storeDir, "metadata"))
-	require.NoError(t, err)
-	assert.Len(t, snapshots, 2,
-		"the new snapshot is on the store beside the planted one")
-}
-
-// writeBackupConfig builds a hermetic config with one small file in its
-// source directory and an empty destination store directory. Returns the
-// config path and the store directory.
-func writeBackupConfig(t *testing.T) (string, string) {
-	t.Helper()
-
-	configPath := writeHermeticPruneConfig(t, false)
-
-	// writeHermeticPruneConfig puts the source and store directories
-	// beside the config file.
-	dir := filepath.Dir(configPath)
-	sourceDir := filepath.Join(dir, "source")
-
-	require.NoError(t, os.Mkdir(sourceDir, 0o750))
-	require.NoError(t, os.WriteFile(
-		filepath.Join(sourceDir, "file.txt"), []byte("hello"), 0o600))
-
-	return configPath, filepath.Join(dir, "store")
-}
@@ -71,11 +71,6 @@ otherwise from the platform config directory (~/.config/vaultik/config.yml
 on Linux), then /etc/vaultik/config.yml.`,
 		Args: cobra.ArbitraryArgs,
 		RunE: func(cmd *cobra.Command, args []string) error {
-			err := opts.Validate()
-			if err != nil {
-				return err
-			}
-
 			// Pass snapshot names from args
 			opts.Snapshots = args
 			// --skip-errors is a global flag on the root command.
@@ -102,13 +97,6 @@ on Linux), then /etc/vaultik/config.yml.`,
 			}, func(v *vaultik.Vaultik) error {
 				return v.CreateSnapshot(opts)
 			}, func(err error) {
-				if errors.Is(err, vaultik.ErrPostBackupPruneFailed) {
-					log.Error("Post-backup prune failed", "error", err)
-					ReportErrorf("Backup completed, but %v", err)
-
-					return
-				}
-
 				log.Error("Snapshot creation failed", "error", err)
 				ReportErrorf("Snapshot creation failed: %v", err)
 			})
@@ -159,16 +159,13 @@ func (r *BlobRepository) UpdateFinished(
 	return nil
 }

-// UpdateUploaded marks a blob as uploaded and records the size of the
-// stored object as its compressed size. That size can differ from the
-// local copy's when an identical blob was already stored with another
-// compression level or other recipients.
+// UpdateUploaded marks a blob as uploaded
 func (r *BlobRepository) UpdateUploaded(
-	ctx context.Context, tx *sql.Tx, id string, storedSize int64,
+	ctx context.Context, tx *sql.Tx, id string,
 ) error {
 	query := `
 		UPDATE blobs 
-		SET uploaded_ts = ?, compressed_size = ?
+		SET uploaded_ts = ?
 		WHERE id = ?
 	`

@@ -176,9 +173,9 @@ func (r *BlobRepository) UpdateUploaded(

 	var err error
 	if tx != nil {
-		_, err = tx.ExecContext(ctx, query, now, storedSize, id)
+		_, err = tx.ExecContext(ctx, query, now, id)
 	} else {
-		_, err = r.db.ExecWithLog(ctx, query, now, storedSize, id)
+		_, err = r.db.ExecWithLog(ctx, query, now, id)
 	}

 	if err != nil {
@@ -122,8 +122,8 @@ func TestBlobRepositoryUpdates(t *testing.T) {
 		t.Errorf("expected compressed size 500, got %d", updated.CompressedSize)
 	}

-	// Test UpdateUploaded, with a stored object larger than the local copy
-	err = repo.UpdateUploaded(ctx, nil, blob.ID.String(), 600)
+	// Test UpdateUploaded
+	err = repo.UpdateUploaded(ctx, nil, blob.ID.String())
 	if err != nil {
 		t.Fatalf("failed to update blob as uploaded: %v", err)
 	}
@@ -141,10 +141,6 @@ func TestBlobRepositoryUpdates(t *testing.T) {
 	if uploaded.UploadedTS.Before(now.Add(-1 * time.Second)) {
 		t.Error("uploaded timestamp should be around test time")
 	}
-
-	if uploaded.CompressedSize != 600 {
-		t.Errorf("expected the stored size 600, got %d", uploaded.CompressedSize)
-	}
 }

 func TestBlobRepositoryDuplicate(t *testing.T) {
@@ -1509,8 +1509,7 @@ func (s *Scanner) finalizeProcessPhase(ctx context.Context, result *ScanResult)
 			// snapshot references has uploaded_ts set, so deduplication and
 			// interrupted-run repair treat these blobs as trustworthy.
 			err = s.repos.WithTx(ctx, func(ctx context.Context, tx *sql.Tx) error {
-				err := s.repos.Blobs.UpdateUploaded(
-					ctx, tx, b.ID, b.Compressed)
+				err := s.repos.Blobs.UpdateUploaded(ctx, tx, b.ID)
 				if err != nil {
 					return fmt.Errorf("marking blob uploaded: %w", err)
 				}
@@ -1546,7 +1545,7 @@ func (s *Scanner) handleBlobReady(
 	blobPath := fmt.Sprintf("blobs/%s/%s/%s",
 		finishedBlob.Hash[:2], finishedBlob.Hash[2:4], finishedBlob.Hash)

-	storedSize, blobExists, err := s.uploadBlobIfNeeded(
+	blobExists, err := s.uploadBlobIfNeeded(
 		ctx, blobPath, blobWithReader, startTime, result)
 	if err != nil {
 		s.cleanupBlobTempFile(blobWithReader)
@@ -1554,8 +1553,7 @@ func (s *Scanner) handleBlobReady(
 		return fmt.Errorf("uploading blob %s: %w", finishedBlob.Hash, err)
 	}

-	err = s.recordBlobMetadata(
-		ctx, finishedBlob, storedSize, blobExists, startTime)
+	err = s.recordBlobMetadata(ctx, finishedBlob, blobExists, startTime)
 	if err != nil {
 		s.cleanupBlobTempFile(blobWithReader)

@@ -1577,22 +1575,20 @@ func (s *Scanner) handleBlobReady(
 }

 // uploadBlobIfNeeded uploads the blob to storage if it doesn't already
-// exist. It returns the size of the object in storage, which for an
-// existing object can differ from the local copy's, and whether it
-// already existed.
+// exist, returns whether it existed
 func (s *Scanner) uploadBlobIfNeeded(
 	ctx context.Context,
 	blobPath string,
 	blobWithReader *blob.WithReader,
 	startTime time.Time,
 	result *ScanResult,
-) (int64, bool, error) {
+) (bool, error) {
 	finishedBlob := blobWithReader.FinishedBlob

 	// Check if blob already exists (deduplication after restart)
 	destination := s.storage.Info().Location

-	stored, err := s.storage.Stat(ctx, blobPath)
+	_, err := s.storage.Stat(ctx, blobPath)
 	if err == nil {
 		log.Info("Blob already exists in storage, skipping upload",
 			"hash", finishedBlob.Hash,
@@ -1601,7 +1597,7 @@ func (s *Scanner) uploadBlobIfNeeded(
 			s.ui.Hex(finishedBlob.Hash), s.ui.Size(finishedBlob.Compressed),
 			s.ui.Path(destination))

-		return stored.Size, true, nil
+		return true, nil
 	}

 	s.ui.Beginf("Uploading blob %s (%s) to %s.",
@@ -1615,7 +1611,7 @@ func (s *Scanner) uploadBlobIfNeeded(
 	if err != nil {
 		log.Error("Failed to upload blob", "hash", finishedBlob.Hash, "error", err)

-		return 0, false, fmt.Errorf("uploading blob to storage: %w", err)
+		return false, fmt.Errorf("uploading blob to storage: %w", err)
 	}

 	uploadDuration := time.Since(startTime)
@@ -1646,7 +1642,7 @@ func (s *Scanner) uploadBlobIfNeeded(
 		stats.BytesUploaded.Add(finishedBlob.Compressed)
 	}

-	return finishedBlob.Compressed, false, nil
+	return false, nil
 }

 // makeUploadProgressCallback creates a progress callback for blob uploads.
@@ -1716,7 +1712,6 @@ func (s *Scanner) makeUploadProgressCallback(
 func (s *Scanner) recordBlobMetadata(
 	ctx context.Context,
 	finishedBlob *blob.FinishedBlob,
-	storedSize int64,
 	blobExists bool,
 	startTime time.Time,
 ) error {
@@ -1728,8 +1723,7 @@ func (s *Scanner) recordBlobMetadata(
 	uploadDuration := time.Since(startTime)

 	return s.repos.WithTx(ctx, func(txCtx context.Context, tx *sql.Tx) error {
-		err := s.repos.Blobs.UpdateUploaded(
-			txCtx, tx, finishedBlob.ID, storedSize)
+		err := s.repos.Blobs.UpdateUploaded(txCtx, tx, finishedBlob.ID)
 		if err != nil {
 			return fmt.Errorf("updating blob upload timestamp: %w", err)
 		}
@@ -1,15 +1,16 @@
 // Package faultstore provides a storage.Storer wrapper that injects
 // faults on demand, so tests can reproduce the failure modes a real
 // backend exhibits: an upload that fails partway, a backend that reports
-// success while storing nothing, and reads that return corrupt or
-// truncated bytes. It is the seam called for by the fault-injection
-// tests (sneak/vaultik issue 72) and is meant to be reused by future
-// tests rather than re-implemented per case.
+// success while storing nothing, reads that return corrupt or truncated
+// bytes, and a delete the backend refuses. It is the seam called for by
+// the fault-injection tests (sneak/vaultik issue 72) and is meant to be
+// reused by future tests rather than re-implemented per case.
 //
-// The wrapper delegates every method to the inner Storer. Two hooks
-// change that: OnPut decides the fate of each write, and OnGet decides
-// how each read's bytes are returned. Both are keyed by the object key,
-// so a test can fault only blobs, only metadata, or a single object.
+// The wrapper delegates every method to the inner Storer. Three hooks
+// change that: OnPut decides the fate of each write, OnGet decides how
+// each read's bytes are returned, and OnDelete decides whether each
+// delete fails. All are keyed by the object key, so a test can fault
+// only blobs, only metadata, or a single object.
 package faultstore

 import (
@@ -25,6 +26,10 @@ import (
 // ErrInjectedUpload is returned by a Put the OnPut hook chose to fail.
 var ErrInjectedUpload = errors.New("faultstore: injected upload failure")

+// ErrInjectedDelete is returned by a Delete the OnDelete hook chose to
+// fail.
+var ErrInjectedDelete = errors.New("faultstore: injected delete failure")
+
 // PutAction is the disposition OnPut assigns to a write.
 type PutAction int

@@ -66,6 +71,11 @@ type Storer struct {
 	// OnGet, when set, is consulted for every Get with the object key
 	// and damages the returned bytes accordingly.
 	OnGet func(key string) GetFault
+
+	// OnDelete, when set, is consulted before every Delete with the
+	// object key. Returning true fails the delete and leaves the object
+	// in place.
+	OnDelete func(key string) bool
 }

 // New wraps inner. inner must be non-nil.
@@ -154,8 +164,12 @@ func (f *Storer) Stat(ctx context.Context, key string) (*storage.ObjectInfo, err
 	return f.inner.Stat(ctx, key)
 }

-// Delete delegates unchanged.
+// Delete removes the object unless OnDelete fails the delete.
 func (f *Storer) Delete(ctx context.Context, key string) error {
+	if f.OnDelete != nil && f.OnDelete(key) {
+		return fmt.Errorf("%w for %q", ErrInjectedDelete, key)
+	}
+
 	return f.inner.Delete(ctx, key)
 }

@@ -167,6 +167,10 @@ const (
 	// metadataKeyParts is the minimum "/"-separated segment count of a
 	// metadata object key (metadata/<snapshot-id>/<filename>).
 	metadataKeyParts = 3
+
+	// blobKeyParts is the minimum "/"-separated segment count of a blob
+	// object key (blobs/<aa>/<bb>/<hash>).
+	blobKeyParts = 4
 )

 // SnapshotMetadataInfo contains information about a single snapshot's metadata
@@ -450,16 +454,23 @@ func (v *Vaultik) scanRemoteBlobStorage(
 		v.stdoutf("Scanning blobs...\n")
 	}

-	// Use prune's listing, so a file under blobs/ that prune would never
-	// delete is not counted as a blob, and so not as an orphaned one.
-	allBlobs, err := v.listAllRemoteBlobs()
-	if err != nil {
-		return err
+	blobCh := v.Storage.ListStream(v.ctx, "blobs/")
+	allBlobs := make(map[string]int64)
+
+	for obj := range blobCh {
+		if obj.Err != nil {
+			return fmt.Errorf("listing blobs: %w", obj.Err)
 		}

-	for _, size := range allBlobs {
+		parts := strings.Split(obj.Key, "/")
+		if len(parts) < blobKeyParts {
+			continue
+		}
+
+		hash := parts[3]
+		allBlobs[hash] = obj.Size
 		result.TotalBlobCount++
-		result.TotalBlobSize += size
+		result.TotalBlobSize += obj.Size
 	}

 	// A blob named only by a manifest that could not be read, or by one
@@ -5,12 +5,14 @@ import (
 	"io/fs"
 	"os"
 	"path/filepath"
+	"strings"
 	"testing"

 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"sneak.berlin/go/vaultik/internal/log"
 	"sneak.berlin/go/vaultik/internal/snapshot"
+	"sneak.berlin/go/vaultik/internal/storage/faultstore"
 )

 // TestNukeRemoteLeavesNoFiles checks that remote nuke leaves no file
@@ -23,7 +25,7 @@ func TestNukeRemoteLeavesNoFiles(t *testing.T) {

 	ctx := context.Background()
 	storeDir := filepath.Join(t.TempDir(), "store")
-	v, repos, _ := backUpToFileDestination(ctx, t, storeDir)
+	v, repos, out := backUpToFileDestination(ctx, t, storeDir)

 	snapshots, err := repos.Snapshots.ListRecent(ctx, listRecentTestLimit)
 	require.NoError(t, err)
@@ -61,4 +63,82 @@ func TestNukeRemoteLeavesNoFiles(t *testing.T) {
 		})
 	require.NoError(t, err)
 	assert.Empty(t, files)
+
+	// Nuke deletes the blobs itself, so advice to run prune is wrong.
+	assert.Contains(t, out.String(), "Backup destination store is now empty.")
+	assert.NotContains(t, out.String(), "vaultik prune")
+}
+
+// TestNukeRemoteFailsWhenADeleteFails checks that remote nuke returns an
+// error when a snapshot's metadata or a blob cannot be deleted, does not
+// report the store empty, and advises running nuke again: prune deletes
+// no metadata.
+func TestNukeRemoteFailsWhenADeleteFails(t *testing.T) {
+	log.Initialize(log.Config{})
+	t.Parallel()
+
+	for _, refusedPrefix := range []string{"metadata/", "blobs/"} {
+		t.Run(refusedPrefix, func(t *testing.T) {
+			t.Parallel()
+
+			ctx := context.Background()
+			storeDir := filepath.Join(t.TempDir(), "store")
+			v, _, out := backUpToFileDestination(ctx, t, storeDir)
+
+			store := faultstore.New(v.Storage)
+			store.OnDelete = func(key string) bool {
+				return strings.HasPrefix(key, refusedPrefix)
+			}
+			v.Storage = store
+
+			require.Error(t, v.NukeRemote(true))
+
+			assert.NotContains(t, out.String(), "now empty")
+			assert.Contains(t, out.String(), "vaultik remote nuke --force")
+			assert.NotContains(t, out.String(), "vaultik prune")
+		})
+	}
+}
+
+// TestNukeRemoteRefusesAnotherDestinationWithoutRetryAdvice checks that
+// remote nuke, when the local index belongs to a different destination
+// than the configured one, fails without advising to run nuke again:
+// that run would fail the same way.
+func TestNukeRemoteRefusesAnotherDestinationWithoutRetryAdvice(t *testing.T) {
+	log.Initialize(log.Config{})
+	t.Parallel()
+
+	ctx := context.Background()
+	storeDir := filepath.Join(t.TempDir(), "store")
+	v, _, out := backUpToFileDestination(ctx, t, storeDir)
+
+	v.Config.StorageURL = "file://" + filepath.Join(t.TempDir(), "elsewhere")
+
+	err := v.NukeRemote(true)
+
+	require.ErrorContains(t, err, "bound to a different backup destination")
+	assert.NotContains(t, out.String(), "vaultik remote nuke --force")
+}
+
+// TestNukeRemoteFailsWhenALocalSnapshotCannotBeRemoved checks that a
+// snapshot remote nuke cannot remove from the local index fails the
+// run: nuke then leaves that snapshot's metadata on the destination
+// store.
+func TestNukeRemoteFailsWhenALocalSnapshotCannotBeRemoved(t *testing.T) {
+	log.Initialize(log.Config{})
+	t.Parallel()
+
+	ctx := context.Background()
+	storeDir := filepath.Join(t.TempDir(), "store")
+	v, _, out := backUpToFileDestination(ctx, t, storeDir)
+
+	_, err := v.DB.Conn().ExecContext(ctx, `
+		CREATE TRIGGER refuse_snapshot_delete BEFORE DELETE ON snapshots
+		BEGIN SELECT RAISE(ABORT, 'simulated index error'); END`)
+	require.NoError(t, err)
+
+	require.Error(t, v.NukeRemote(true))
+
+	assert.NotContains(t, out.String(), "now empty")
+	assert.Contains(t, out.String(), "vaultik remote nuke --force")
 }
@@ -19,14 +19,18 @@ type PruneOptions struct {
 var errNukeRequiresForce = errors.New(
 	"nuke requires --force (this deletes ALL remote snapshots and blobs)")

+// errBlobDeleteFailed reports unreferenced blobs prune could not delete.
+var errBlobDeleteFailed = errors.New("blob(s) could not be deleted")
+
+// nukeCommandHint is the command suggested when a remote nuke could not
+// delete everything: running it again deletes what is left, which
+// `vaultik prune` does not do for snapshot metadata.
+const nukeCommandHint = "vaultik remote nuke --force"
+
 // metadataDirName is the top-level remote directory holding snapshot
 // metadata.
 const metadataDirName = "metadata"

-// blobKeyParts is the "/"-separated segment count of a blob object key
-// (blobs/<aa>/<bb>/<hash>).
-const blobKeyParts = 4
-
 // NukeRemote deletes every snapshot's metadata and every blob from remote
 // storage, along with any object an upload cut off part-way left under a
 // temporary `.partial` name. After this returns successfully the bucket
@@ -39,6 +43,29 @@ func (v *Vaultik) NukeRemote(force bool) error {
 		return errNukeRequiresForce
 	}

+	// Checked before anything is deleted: running nuke again would fail
+	// the same way, so the advice below must not follow this error.
+	err := v.EnsureStorageBinding()
+	if err != nil {
+		return err
+	}
+
+	err = v.emptyDestinationStore()
+	if err != nil {
+		v.UI.Warningf("Run '%s' again to delete what is left on the "+
+			"backup destination store.", nukeCommandHint)
+
+		return err
+	}
+
+	v.UI.Completef("Backup destination store is now empty.")
+
+	return nil
+}
+
+// emptyDestinationStore does the deleting for NukeRemote and stops at
+// the first step that fails.
+func (v *Vaultik) emptyDestinationStore() error {
 	v.UI.Beginf("Removing all snapshot metadata from backup destination store.")

 	_, err := v.RemoveAllSnapshots(&RemoveOptions{Force: true})
@@ -62,8 +89,6 @@ func (v *Vaultik) NukeRemote(force bool) error {
 		}
 	}

-	v.UI.Completef("Backup destination store is now empty.")
-
 	return nil
 }

@@ -166,14 +191,17 @@ func (v *Vaultik) PruneBlobs(opts *PruneOptions) error {
 	v.deleteUnreferencedBlobs(unreferencedBlobs, allBlobs, result)

 	if opts.JSON {
-		return v.outputPruneBlobsJSON(result)
+		err = v.outputPruneBlobsJSON(result)
+		if err != nil {
+			return err
 		}
-
+	} else {
 		v.stdoutf("\nDeleted %d blob(s) totaling %s\n",
 			result.BlobsDeleted, ubytes(result.BytesFreed))
+	}

 	if result.BlobsFailed > 0 {
-		v.stdoutf("Failed to delete %d blob(s)\n", result.BlobsFailed)
+		return fmt.Errorf("%d %w", result.BlobsFailed, errBlobDeleteFailed)
 	}

 	return nil
@@ -0,0 +1,98 @@
+package vaultik_test
+
+import (
+	"context"
+	"encoding/json"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"sneak.berlin/go/vaultik/internal/log"
+	"sneak.berlin/go/vaultik/internal/storage/faultstore"
+	"sneak.berlin/go/vaultik/internal/vaultik"
+)
+
+// refuseBlobDeletes is an OnDelete hook that fails every blob delete.
+func refuseBlobDeletes(key string) bool {
+	return strings.HasPrefix(key, "blobs/")
+}
+
+// TestPruneBlobs_FailedDeleteReturnsError checks that prune returns an
+// error, so the command exits non-zero, when an unreferenced blob cannot
+// be deleted.
+func TestPruneBlobs_FailedDeleteReturnsError(t *testing.T) {
+	log.Initialize(log.Config{})
+	t.Parallel()
+
+	env := newListEnv(t)
+
+	// The store holds no manifest, so nothing references this blob.
+	addBlob(t, env.store.testStorer, testBlobHashA)
+
+	store := faultstore.New(env.store)
+	store.OnDelete = refuseBlobDeletes
+	env.v.Storage = store
+
+	err := env.v.PruneBlobs(&vaultik.PruneOptions{Force: true})
+
+	require.ErrorContains(t, err, "1 blob(s) could not be deleted")
+}
+
+// TestPruneBlobs_JSONFailedDeleteReturnsError checks that prune --json
+// prints its document and then returns an error when an unreferenced
+// blob cannot be deleted.
+func TestPruneBlobs_JSONFailedDeleteReturnsError(t *testing.T) {
+	log.Initialize(log.Config{})
+	t.Parallel()
+
+	env := newListEnv(t)
+	addBlob(t, env.store.testStorer, testBlobHashA)
+
+	store := faultstore.New(env.store)
+	store.OnDelete = refuseBlobDeletes
+	env.v.Storage = store
+
+	err := env.v.PruneBlobs(&vaultik.PruneOptions{JSON: true})
+
+	require.Error(t, err, "prune must fail when a blob cannot be deleted")
+
+	var result vaultik.PruneBlobsResult
+
+	require.NoError(t, json.Unmarshal(env.stdout.Bytes(), &result),
+		"stdout must hold the JSON document, got:\n%s", env.stdout.String())
+	assert.Equal(t, 1, result.BlobsFailed)
+	assert.Equal(t, 0, result.BlobsDeleted)
+}
+
+// TestCreateSnapshotPrune_FailedBlobDeleteReturnsError checks that the
+// prune step of snapshot create --prune fails the run when an
+// unreferenced blob cannot be deleted.
+func TestCreateSnapshotPrune_FailedBlobDeleteReturnsError(t *testing.T) {
+	log.Initialize(log.Config{})
+	t.Parallel()
+
+	ctx := context.Background()
+	storeDir := filepath.Join(t.TempDir(), "store")
+	v, _, _ := backUpToFileDestination(ctx, t, storeDir)
+
+	// No snapshot references this blob, so the prune step deletes it.
+	orphan := filepath.Join(storeDir, "blobs", testBlobHashA[:2],
+		testBlobHashA[2:4], testBlobHashA)
+	require.NoError(t, os.MkdirAll(filepath.Dir(orphan), 0o750))
+	require.NoError(t, os.WriteFile(orphan, []byte("orphan"), 0o600))
+
+	store := faultstore.New(v.Storage)
+	store.OnDelete = refuseBlobDeletes
+	v.Storage = store
+
+	err := v.CreateSnapshot(&vaultik.SnapshotCreateOptions{
+		Cron:  true,
+		Prune: true,
+	})
+
+	require.ErrorContains(t, err, "1 blob(s) could not be deleted")
+	assert.FileExists(t, orphan)
+}
@@ -185,39 +185,6 @@ func TestRemoteInfo_SkipsNonConformingMetadataName(t *testing.T) {
 	assert.NotContains(t, doc, "unreadable_manifests")
 }

-// TestRemoteInfo_DoesNotCountNonBlobFileUnderBlobs checks that a file
-// under blobs/ whose name is not a blob hash is not counted as a blob.
-// prune never deletes such a file, so counting it as an orphaned blob
-// would advise a prune that removes nothing.
-func TestRemoteInfo_DoesNotCountNonBlobFileUnderBlobs(t *testing.T) {
-	log.Initialize(log.Config{})
-	t.Parallel()
-
-	env := newListEnv(t)
-	env.addRemote(t, listRemoteID, time.Date(2026, 3, 2, 0, 0, 0, 0, time.UTC))
-	addBlob(t, env.store.testStorer, testBlobHashA)
-	require.NoError(t, env.store.Put(context.Background(),
-		"blobs/"+testBlobHashA[:2]+"/"+testBlobHashA[2:4]+"/.DS_Store",
-		bytes.NewReader([]byte("junk"))))
-
-	require.NoError(t, env.v.RemoteInfo(false))
-
-	text := env.stdout.String()
-	assert.Contains(t, text, "Total blobs on remote:      1 (")
-	assert.Contains(t, text, "Orphaned (unreferenced):    0 (")
-	assert.NotContains(t, text, "vaultik prune")
-
-	env.stdout.Reset()
-	require.NoError(t, env.v.RemoteInfo(true))
-
-	var doc map[string]any
-
-	require.NoError(t, json.Unmarshal(env.stdout.Bytes(), &doc))
-	assert.InDelta(t, 1, doc["total_blob_count"], 0)
-	assert.InDelta(t, 0, doc["orphaned_blob_count"], 0)
-	assert.InDelta(t, 0, doc["orphaned_blob_size"], 0)
-}
-
 // TestRemoteInfo_DirectoryWithoutManifestLeavesOrphansKnown checks that
 // a directory under metadata/ holding no manifest.json.zst, such as one
 // left by a backup interrupted before its manifest upload, leaves the
@@ -476,7 +476,8 @@ func TestRemoveAllSnapshots_WithForce(t *testing.T) {
 	out := tv.Stdout.String()
 	assert.Contains(t, out, "Removed 2 snapshot(s)")
 	assert.Contains(t, out, "Removed snapshot metadata from remote storage")
-	assert.Contains(t, out, "vaultik prune")
+	// Remote nuke, the only caller, deletes the blobs next itself.
+	assert.NotContains(t, out, "vaultik prune")
 }

 func TestRemoveAllSnapshots_DryRun(t *testing.T) {
@@ -1,83 +0,0 @@
-package vaultik_test
-
-import (
-	"context"
-	"io"
-	"path/filepath"
-	"testing"
-
-	"filippo.io/age"
-	"github.com/spf13/afero"
-	"github.com/stretchr/testify/require"
-	"sneak.berlin/go/vaultik/internal/log"
-	"sneak.berlin/go/vaultik/internal/ui"
-	"sneak.berlin/go/vaultik/internal/vaultik"
-)
-
-// TestVerifyPassesWhenBackupReusesStoredBlob backs up a file, then backs it
-// up again from an empty local index, as after `database delete`, with an
-// extra age recipient. The second backup builds the same blob, finds it
-// already stored and skips the upload, but its own copy is larger, because
-// age writes a header entry for each recipient. The second snapshot must
-// record the size of the stored object, so shallow and deep verify pass.
-func TestVerifyPassesWhenBackupReusesStoredBlob(t *testing.T) {
-	log.Initialize(log.Config{})
-	t.Parallel()
-
-	fs := afero.NewOsFs()
-	tempDir := t.TempDir()
-
-	dataDir := filepath.Join(tempDir, "source")
-	storeDir := filepath.Join(tempDir, "remote")
-	secondDBPath := filepath.Join(tempDir, "second.sqlite")
-
-	chunkSize := int64(64 * 1024)
-	maxBlobSize := int64(512 * 1024)
-
-	require.NoError(t, fs.MkdirAll(dataDir, 0o755))
-	require.NoError(t, afero.WriteFile(fs,
-		filepath.Join(dataDir, "data.bin"),
-		bytesPattern("reuse-", int(chunkSize*3)), 0o644))
-
-	ctx := context.Background()
-
-	runFileStorageBackup(ctx, t, fs, dataDir, storeDir,
-		filepath.Join(tempDir, "first.sqlite"), chunkSize, maxBlobSize)
-
-	env := setupDedupBackupEnv(
-		ctx, t, fs, storeDir, secondDBPath, chunkSize, maxBlobSize)
-
-	defer func() { _ = env.db.Close() }()
-
-	extra, err := age.GenerateX25519Identity()
-	require.NoError(t, err)
-
-	env.cfg.AgeRecipients = append(env.cfg.AgeRecipients,
-		extra.Recipient().String())
-
-	snapshotID, result := runDedupSnapshot(ctx, t, env.sm, env.makeScanner(),
-		env.cfg.Hostname, "second", dataDir, secondDBPath)
-	require.Positive(t, result.BlobsCreated)
-	require.Zero(t, result.BlobsUploaded,
-		"the second backup should find every blob already stored")
-
-	require.NoError(t, env.db.Close())
-
-	verifier := &vaultik.Vaultik{
-		Config:  env.cfg,
-		Storage: env.storer,
-		Fs:      fs,
-		Stdout:  io.Discard,
-		Stderr:  io.Discard,
-		UI:      ui.NewWithColor(io.Discard, false),
-	}
-	verifier.SetContext(ctx)
-
-	require.NoError(t,
-		verifier.VerifySnapshotWithOptions(snapshotID, &vaultik.VerifyOptions{}),
-		"shallow verify should pass")
-	require.NoError(t,
-		verifier.VerifySnapshotWithOptions(snapshotID,
-			&vaultik.VerifyOptions{Deep: true}),
-		"deep verify should pass")
-}
@@ -24,18 +24,13 @@ var (
 	errSnapshotVerifyFailed = errors.New("verification failed")
 	errRemoveAllNeedsForce  = errors.New("--all requires --force")
 	errSnapshotNotFound     = errors.New("snapshot not found")
+	errSnapshotRemoveFailed = errors.New("snapshot(s) could not be removed")
 	errInvalidTableName     = errors.New("invalid table name")
 	errNoAgeRecipients      = errors.New(
 		"creating a snapshot needs at least one public key in " +
 			"age_recipients (generate a keypair with: age-keygen)")
-	errKeepNewerThanNeedsPrune = errors.New("--keep-newer-than requires --prune")
 )

-// ErrPostBackupPruneFailed wraps a failure of the post-backup prune,
-// which runs only after every snapshot was created and uploaded, so the
-// caller can report it apart from a failed backup.
-var ErrPostBackupPruneFailed = errors.New("post-backup prune failed")
-
 // listRecentLimit caps how many snapshot rows are fetched from the
 // local index for listing and purge operations.
 const listRecentLimit = 10000
@@ -51,26 +46,6 @@ type SnapshotCreateOptions struct {
 	Snapshots     []string // Optional list of snapshot names to process (empty = all)
 }

-// Validate fails when KeepNewerThan is set without Prune or is not a
-// valid duration. Only the post-backup prune reads KeepNewerThan, after
-// the backup, so the caller checks it before starting one.
-func (opts *SnapshotCreateOptions) Validate() error {
-	if opts.KeepNewerThan == "" {
-		return nil
-	}
-
-	if !opts.Prune {
-		return errKeepNewerThanNeedsPrune
-	}
-
-	_, err := parseDuration(opts.KeepNewerThan)
-	if err != nil {
-		return fmt.Errorf("--keep-newer-than: %w", err)
-	}
-
-	return nil
-}
-
 // CreateSnapshot executes the snapshot creation operation
 func (v *Vaultik) CreateSnapshot(opts *SnapshotCreateOptions) error {
 	// config.Load accepts an empty list, since listing, verifying and
@@ -149,7 +124,7 @@ func (v *Vaultik) CreateSnapshot(opts *SnapshotCreateOptions) error {
 	if opts.Prune {
 		err = v.runPostBackupPrune(snapshotNames, opts.KeepNewerThan)
 		if err != nil {
-			return fmt.Errorf("%w: %w", ErrPostBackupPruneFailed, err)
+			return fmt.Errorf("post-backup prune: %w", err)
 		}
 	}

@@ -1539,8 +1514,8 @@ func (v *Vaultik) handleRemoveAllDryRun(

 // executeRemoveAll deletes every local snapshot and, unless LocalOnly
 // is set, every corresponding remote metadata directory plus any
-// orphan remote keys. Blobs are not touched; the printed prune-command
-// hint is the next step.
+// orphan remote keys. Blobs are not touched. It returns an error when
+// any snapshot could not be removed.
 func (v *Vaultik) executeRemoveAll(
 	localSnaps, orphanRemoteKeys []string, opts *RemoveOptions,
 ) (*RemoveResult, error) {
@@ -1554,18 +1529,11 @@ func (v *Vaultik) executeRemoveAll(

 	result := &RemoveResult{}

-	remoteErrors := v.removeAllLocalSnapshots(localSnaps, opts, result)
+	failures := v.removeAllLocalSnapshots(localSnaps, opts, result)

 	if !opts.LocalOnly {
-		remoteErrors += v.removeAllOrphanRemotes(orphanRemoteKeys, result)
-
-		if remoteErrors == 0 {
-			result.RemoteRemoved = true
-		} else if v.UI != nil {
-			v.UI.Warningf("Some remote metadata deletions failed. Run '%s' "+
-				"once the remote is healthy to clean up unreferenced blobs.",
-				pruneCommandHint)
-		}
+		failures += v.removeAllOrphanRemotes(orphanRemoteKeys, result)
+		result.RemoteRemoved = failures == 0
 	}

 	if v.SnapshotManager != nil {
@@ -1575,6 +1543,10 @@ func (v *Vaultik) executeRemoveAll(
 		}
 	}

+	if failures > 0 {
+		return result, fmt.Errorf("%d %w", failures, errSnapshotRemoveFailed)
+	}
+
 	if opts.JSON {
 		return result, v.outputRemoveJSON(result)
 	}
@@ -1583,9 +1555,6 @@ func (v *Vaultik) executeRemoveAll(

 	if !opts.LocalOnly && result.RemoteRemoved {
 		v.printlnStdout("Removed snapshot metadata from remote storage")
-		v.stdoutf("\nNote: Removed snapshots' blobs remain on the remote. "+
-			"Run '%s' to delete any blobs no longer referenced by any "+
-			"remaining remote snapshot.\n", pruneCommandHint)
 	}

 	return result, nil
@@ -1628,12 +1597,13 @@ func (v *Vaultik) deleteIncompleteSnapshot(snapshotIDStr string, id any) bool {
 }

 // removeAllLocalSnapshots removes each known local snapshot (and, unless
-// --local-only, its remote metadata), returning the number of remote
-// deletion failures.
+// --local-only, its remote metadata), returning the number of snapshots
+// it could not remove. A snapshot left in the local index keeps its
+// remote metadata too, since it is not an orphan remote key.
 func (v *Vaultik) removeAllLocalSnapshots(
 	localSnaps []string, opts *RemoveOptions, result *RemoveResult,
 ) int {
-	remoteErrors := 0
+	failures := 0

 	for _, snapshotID := range localSnaps {
 		log.Info("Removing snapshot", "snapshot_id", snapshotID)
@@ -1643,6 +1613,8 @@ func (v *Vaultik) removeAllLocalSnapshots(
 			log.Error("Failed to remove from local database",
 				"snapshot_id", snapshotID, "error", err)

+			failures++
+
 			continue
 		}

@@ -1653,14 +1625,14 @@ func (v *Vaultik) removeAllLocalSnapshots(
 				log.Warn("Failed to remove snapshot metadata from remote",
 					"snapshot_id", snapshotID, "error", err)

-				remoteErrors++
+				failures++
 			}
 		}

 		result.SnapshotsRemoved = append(result.SnapshotsRemoved, snapshotID)
 	}

-	return remoteErrors
+	return failures
 }

 // removeAllOrphanRemotes deletes remote snapshot metadata that has no