Return errors instead of exiting on rollback, rows close and blob temp file cleanup (closes #345)

WithReadTx ended the process when its rollback failed, including after
Ctrl-C, when database/sql has already rolled the transaction back and
Rollback returns sql.ErrTxDone. It now handles the rollback as WithTx
does. The query functions ended the process when closing their rows
failed; they now ignore the deferred close, because database/sql reports
a close failure through rows.Err(), which each returns, once the rows
are read to the end. The unused database.Fatalf is removed.
cleanupBlobTempFile returns its close or remove failure and the upload
step returns it, so such a failure now fails the backup.

No test makes a rows close fail: SQLite fails it only with the error the
last read already returned.

Model: opus-5-5

README.md

@@ -293,8 +293,8 @@ needed.
 **`snapshot list`**: Show every snapshot known to this host — the union
 of the local index and the backup destination store — with timestamps
 and three sizes per snapshot (compressed remote size; total
-uncompressed chunk size; size of the chunks that snapshot stored that
-were not stored before, recorded when its backup ran).
+uncompressed chunk size; size of chunks newly referenced by that
+snapshot).
 
 Listing the destination store does **not** require the age secret key,
 so it works in vaultik's intended configuration, where the backed-up
@@ -307,8 +307,8 @@ only the hash, so hostname and snapshot name exist solely in the local
 index and in the encrypted per-snapshot database. Snapshots found only
 on the destination store are therefore listed as
 `<remote only:<abbreviated remote key>>` and show `<remote only>` in
-the uncompressed and "new chunk" columns, which come only from the
-local index. Their timestamp and compressed size are real,
+the uncompressed and "new chunk" columns, which can only be computed
+from the local index. Their timestamp and compressed size are real,
 read from the manifest.
 
 Snapshots in the local index with no counterpart on the destination

docs/DATAMODEL.md

@@ -119,7 +119,6 @@ Tracks backup snapshots.
 - `completed_at` (INTEGER) - Completion timestamp (NULL if in progress)
 - `file_count` (INTEGER) - Number of files in snapshot
 - `chunk_count` (INTEGER) - Number of chunks this snapshot stored that were not stored before
-- `new_chunk_size` (INTEGER) - Total size of those chunks before compression; `snapshot list` shows it as NEW CHUNK SIZE
 - `blob_count` (INTEGER) - Number of blobs this snapshot created
 - `total_size` (INTEGER) - Total size of all files
 - `blob_size` (INTEGER) - Total compressed size of all referenced blobs

internal/database/blob_chunks.go

@@ -58,12 +58,7 @@ func (r *BlobChunkRepository) GetByBlobID(
 		return nil, fmt.Errorf("querying blob chunks: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var blobChunks []*BlobChunk
 

internal/database/blobs.go

@@ -83,12 +83,7 @@ func (r *BlobRepository) GetAll(ctx context.Context) (map[string]*Blob, error) {
 		return nil, fmt.Errorf("querying blobs: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	out := make(map[string]*Blob)
 

internal/database/chunk_files.go

@@ -61,12 +61,7 @@ func (r *ChunkFileRepository) GetByChunkHash(
 		return nil, fmt.Errorf("querying chunk files: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	return r.scanChunkFiles(rows)
 }
@@ -87,12 +82,7 @@ func (r *ChunkFileRepository) GetByFilePath(
 		return nil, fmt.Errorf("querying chunk files: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	return r.scanChunkFiles(rows)
 }
@@ -112,12 +102,7 @@ func (r *ChunkFileRepository) GetByFileID(
 		return nil, fmt.Errorf("querying chunk files: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	return r.scanChunkFiles(rows)
 }

internal/database/chunks.go

@@ -107,12 +107,7 @@ func (r *ChunkRepository) GetByHashes(
 		return nil, fmt.Errorf("querying chunks: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var chunks []*Chunk
 
@@ -152,12 +147,7 @@ func (r *ChunkRepository) ListUnpacked(
 		return nil, fmt.Errorf("querying unpacked chunks: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var chunks []*Chunk
 

internal/database/chunks_ext.go

@@ -38,12 +38,7 @@ func (r *ChunkRepository) list(ctx context.Context, query string) ([]*Chunk, err
 		return nil, fmt.Errorf("querying chunks: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var chunks []*Chunk
 

internal/database/errors.go

@@ -1,12 +0,0 @@
-package database
-
-import (
-	"fmt"
-	"os"
-)
-
-// Fatalf prints an error message to stderr and exits with status 1
-func Fatalf(format string, args ...any) {
-	fmt.Fprintf(os.Stderr, "FATAL: "+format+"\n", args...)
-	os.Exit(1)
-}

internal/database/file_chunks.go

@@ -62,12 +62,7 @@ func (r *FileChunkRepository) GetByPath(
 		return nil, fmt.Errorf("querying file chunks: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	return r.scanFileChunks(rows)
 }
@@ -88,12 +83,7 @@ func (r *FileChunkRepository) GetByFileID(
 		return nil, fmt.Errorf("querying file chunks: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	return r.scanFileChunks(rows)
 }
@@ -117,12 +107,7 @@ func (r *FileChunkRepository) GetByPathTx(
 		return nil, fmt.Errorf("querying file chunks: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	fileChunks, err := r.scanFileChunks(rows)
 	LogSQL("GetByPathTx", "Complete", path, "count", len(fileChunks))

internal/database/files.go

@@ -175,12 +175,7 @@ func (r *FileRepository) ListModifiedSince(
 		return nil, fmt.Errorf("querying files: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var files []*File
 
@@ -256,12 +251,7 @@ func (r *FileRepository) ListUnderPath(
 		return nil, fmt.Errorf("querying files: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var files []*File
 
@@ -304,12 +294,7 @@ func (r *FileRepository) ListIDsWithChunksNotInUploadedBlobs(
 		return nil, fmt.Errorf("querying files: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var ids []types.FileID
 
@@ -340,12 +325,7 @@ func (r *FileRepository) ListAll(ctx context.Context) ([]*File, error) {
 		return nil, fmt.Errorf("querying files: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var files []*File
 

internal/database/repositories.go

@@ -106,22 +106,24 @@ func (r *Repositories) WithReadTx(ctx context.Context, fn TxFunc) error {
 
 	defer func() {
 		if p := recover(); p != nil {
-			rollbackErr := tx.Rollback()
-			if rollbackErr != nil {
-				Fatalf("failed to rollback transaction: %v", rollbackErr)
-			}
+			// The panic is the failure to report, so a rollback error
+			// is dropped.
+			_ = tx.Rollback()
 
 			panic(p)
-		} else if err != nil {
-			rollbackErr := tx.Rollback()
-			if rollbackErr != nil {
-				Fatalf("failed to rollback transaction: %v", rollbackErr)
-			}
 		}
 	}()
 
 	err = fn(ctx, tx)
 	if err != nil {
+		// Once ctx is cancelled, database/sql rolls the transaction
+		// back itself and Rollback returns sql.ErrTxDone.
+		rollbackErr := tx.Rollback()
+		if rollbackErr != nil && !errors.Is(rollbackErr, sql.ErrTxDone) {
+			return fmt.Errorf("%w; rolling back transaction: %w",
+				err, rollbackErr)
+		}
+
 		return err
 	}
 

internal/database/repositories_test.go

@@ -411,3 +411,41 @@ func TestRepositoriesReadTransaction(t *testing.T) {
 		t.Error("expected to retrieve file in read transaction")
 	}
 }
+
+// When the context is cancelled during a read transaction's body,
+// database/sql rolls the transaction back itself, so WithReadTx's own
+// rollback finds it already done. WithReadTx must return the body's error
+// alone, without that rollback.
+func TestRepositoriesReadTransactionContextCancelled(t *testing.T) {
+	t.Parallel()
+
+	db, cleanup := setupTestDB(t)
+	defer cleanup()
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+
+	repos := database.NewRepositories(db)
+
+	err := repos.WithReadTx(ctx, func(ctx context.Context, tx *sql.Tx) error {
+		cancel()
+
+		// database/sql rolls back from a goroutine of its own; wait
+		// until it has, so the test does not depend on that race.
+		for {
+			_, err := tx.ExecContext(context.WithoutCancel(ctx), "SELECT 1")
+			if errors.Is(err, sql.ErrTxDone) {
+				return ctx.Err()
+			}
+
+			time.Sleep(time.Millisecond)
+		}
+	})
+	if !errors.Is(err, context.Canceled) {
+		t.Fatalf("expected context.Canceled, got: %v", err)
+	}
+
+	if errors.Is(err, sql.ErrTxDone) {
+		t.Fatalf("the already-done rollback was reported: %v", err)
+	}
+}

internal/database/schema/001.sql

@@ -86,7 +86,6 @@ CREATE TABLE IF NOT EXISTS snapshots (
     completed_at INTEGER,
     file_count INTEGER NOT NULL DEFAULT 0,
     chunk_count INTEGER NOT NULL DEFAULT 0,
-    new_chunk_size INTEGER NOT NULL DEFAULT 0,
     blob_count INTEGER NOT NULL DEFAULT 0,
     total_size INTEGER NOT NULL DEFAULT 0,
     blob_size INTEGER NOT NULL DEFAULT 0,

internal/database/snapshots.go

@@ -129,7 +129,6 @@ func (r *SnapshotRepository) UpdateExtendedStats(
 	compressionLevel int,
 	uploadBytes int64,
 	uploadDurationMs int64,
-	newChunkSize int64,
 ) error {
 	compressionRatio, err := r.extendedCompressionRatio(
 		ctx, tx, snapshotID, blobUncompressedSize,
@@ -144,19 +143,18 @@ func (r *SnapshotRepository) UpdateExtendedStats(
 		    compression_ratio = ?,
 		    compression_level = ?,
 		    upload_bytes = ?,
-		    upload_duration_ms = ?,
-		    new_chunk_size = ?
+		    upload_duration_ms = ?
 		WHERE id = ?
 	`
 
 	if tx != nil {
 		_, err = tx.ExecContext(ctx, query,
 			blobUncompressedSize, compressionRatio, compressionLevel,
-			uploadBytes, uploadDurationMs, newChunkSize, snapshotID)
+			uploadBytes, uploadDurationMs, snapshotID)
 	} else {
 		_, err = r.db.ExecWithLog(ctx, query,
 			blobUncompressedSize, compressionRatio, compressionLevel,
-			uploadBytes, uploadDurationMs, newChunkSize, snapshotID)
+			uploadBytes, uploadDurationMs, snapshotID)
 	}
 
 	if err != nil {
@@ -241,12 +239,7 @@ func (r *SnapshotRepository) GetOnlySnapshot(ctx context.Context) (*Snapshot, er
 		return nil, fmt.Errorf("querying snapshots: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	snapshots, err := r.scanSnapshotRows(rows)
 	if err != nil {
@@ -281,12 +274,7 @@ func (r *SnapshotRepository) ListRecent(
 		return nil, fmt.Errorf("querying snapshots: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	return r.scanSnapshotRows(rows)
 }
@@ -501,12 +489,7 @@ func (r *SnapshotRepository) GetBlobHashes(
 		return nil, fmt.Errorf("querying blob hashes: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	var blobs []string
 
@@ -596,24 +579,45 @@ func (r *SnapshotRepository) GetSnapshotUncompressedChunkSize(
 	return totalSize, nil
 }
 
-// GetSnapshotNewChunkSize returns the total plaintext size of the chunks the
-// snapshot's backup stored that were not stored before, as the backup
-// recorded it. It is not derived from the snapshot's files: a file's row is
-// shared by every snapshot that lists it and is rewritten when the file
-// changes, so it shows only the file's latest chunks.
+// GetSnapshotNewChunkSize returns the sum of plaintext sizes of chunks that are
+// referenced by this snapshot but not by any earlier completed snapshot known to
+// the local database. The result is the marginal uncompressed data this snapshot
+// added to the dedup pool — i.e., the delta from prior snapshots.
 func (r *SnapshotRepository) GetSnapshotNewChunkSize(
 	ctx context.Context, snapshotID string,
 ) (int64, error) {
-	query := `SELECT new_chunk_size FROM snapshots WHERE id = ?`
+	query := `
+		WITH this_snap_chunks AS (
+			SELECT DISTINCT fc.chunk_hash
+			FROM snapshot_files sf
+			JOIN file_chunks fc ON sf.file_id = fc.file_id
+			WHERE sf.snapshot_id = ?
+		),
+		prior_chunks AS (
+			SELECT DISTINCT fc.chunk_hash
+			FROM snapshots s
+			JOIN snapshot_files sf ON sf.snapshot_id = s.id
+			JOIN file_chunks fc ON fc.file_id = sf.file_id
+			WHERE s.completed_at IS NOT NULL
+			  AND s.id != ?
+			  AND s.started_at < (SELECT started_at FROM snapshots WHERE id = ?)
+		)
+		SELECT COALESCE(SUM(c.size), 0)
+		FROM chunks c
+		JOIN this_snap_chunks t ON c.chunk_hash = t.chunk_hash
+		WHERE c.chunk_hash NOT IN (SELECT chunk_hash FROM prior_chunks)
+	`
 
-	var newChunkSize int64
+	var totalSize int64
 
-	err := r.db.conn.QueryRowContext(ctx, query, snapshotID).Scan(&newChunkSize)
+	err := r.db.conn.QueryRowContext(
+		ctx, query, snapshotID, snapshotID, snapshotID,
+	).Scan(&totalSize)
 	if err != nil {
 		return 0, fmt.Errorf("querying new chunk size: %w", err)
 	}
 
-	return newChunkSize, nil
+	return totalSize, nil
 }
 
 // GetIncompleteSnapshots returns all snapshots that haven't been completed
@@ -634,12 +638,7 @@ func (r *SnapshotRepository) GetIncompleteSnapshots(
 		return nil, fmt.Errorf("querying incomplete snapshots: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	return r.scanSnapshotRows(rows)
 }
@@ -662,12 +661,7 @@ func (r *SnapshotRepository) GetIncompleteByHostname(
 		return nil, fmt.Errorf("querying incomplete snapshots: %w", err)
 	}
 
-	defer func() {
-		err := rows.Close()
-		if err != nil {
-			Fatalf("failed to close rows: %v", err)
-		}
-	}()
+	defer func() { _ = rows.Close() }()
 
 	// Same column set as every other multi-row snapshot query, so the
 	// shared scanner applies — including its timestamp normalization.

internal/snapshot/blob_temp_file_test.go

@@ -0,0 +1,62 @@
+//nolint:testpackage // exercises the unexported cleanupBlobTempFile
+package snapshot
+
+import (
+	"errors"
+	"os"
+	"testing"
+
+	"github.com/spf13/afero"
+	"sneak.berlin/go/vaultik/internal/blob"
+)
+
+// A blob temp file that cannot be closed or removed must make the cleanup
+// return an error, not end the process.
+func TestCleanupBlobTempFileReturnsErrors(t *testing.T) {
+	t.Parallel()
+
+	fs := afero.NewOsFs()
+	s := &Scanner{fs: fs}
+
+	tests := []struct {
+		name string
+		// prepare runs before the cleanup and makes its close or remove
+		// fail.
+		prepare func(tempFile afero.File) error
+		want    error
+	}{
+		{
+			name:    "close fails",
+			prepare: func(tempFile afero.File) error { return tempFile.Close() },
+			want:    os.ErrClosed,
+		},
+		{
+			name: "remove fails",
+			prepare: func(tempFile afero.File) error {
+				return fs.Remove(tempFile.Name())
+			},
+			want: os.ErrNotExist,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+
+			tempFile, err := afero.TempFile(fs, t.TempDir(), "blob-*.tmp")
+			if err != nil {
+				t.Fatal(err)
+			}
+
+			err = tt.prepare(tempFile)
+			if err != nil {
+				t.Fatal(err)
+			}
+
+			err = s.cleanupBlobTempFile(&blob.WithReader{TempFile: tempFile})
+			if !errors.Is(err, tt.want) {
+				t.Fatalf("expected an error wrapping %v, got: %v", tt.want, err)
+			}
+		})
+	}
+}

internal/snapshot/scanner.go

@@ -135,9 +135,7 @@ type ScannerConfig struct {
 // are counted per file: BytesScanned is the size of the new and changed
 // files, BytesSkipped that of the unchanged ones. FilesFailed counts the
 // new and changed files that phase 2 could not store; FilesScanned
-// includes them and BytesScanned does not. ChunksCreated counts the
-// chunks that were not already stored, and NewChunkSize is their total
-// size.
+// includes them and BytesScanned does not.
 type ScanResult struct {
 	FilesScanned   int
 	FilesSkipped   int
@@ -147,7 +145,6 @@ type ScanResult struct {
 	BytesSkipped   int64
 	BytesDeleted   int64
 	ChunksCreated  int
-	NewChunkSize   int64
 	BlobsCreated   int
 	BlobsUploaded  int
 	BytesUploaded  int64
@@ -1556,19 +1553,23 @@ func (s *Scanner) handleBlobReady(
 	blobExists, err := s.uploadBlobIfNeeded(
 		ctx, blobPath, blobWithReader, startTime, result)
 	if err != nil {
-		s.cleanupBlobTempFile(blobWithReader)
+		cleanupErr := s.cleanupBlobTempFile(blobWithReader)
 
-		return fmt.Errorf("uploading blob %s: %w", finishedBlob.Hash, err)
+		return errors.Join(
+			fmt.Errorf("uploading blob %s: %w", finishedBlob.Hash, err), cleanupErr)
 	}
 
 	err = s.recordBlobMetadata(ctx, finishedBlob, blobExists, startTime)
 	if err != nil {
-		s.cleanupBlobTempFile(blobWithReader)
+		cleanupErr := s.cleanupBlobTempFile(blobWithReader)
 
-		return err
+		return errors.Join(err, cleanupErr)
 	}
 
-	s.cleanupBlobTempFile(blobWithReader)
+	err = s.cleanupBlobTempFile(blobWithReader)
+	if err != nil {
+		return err
+	}
 
 	// Chunks from this blob are now committed to DB - remove from pending set
 	s.removePendingChunkHashes(blobWithReader.InsertedChunkHashes)
@@ -1762,20 +1763,22 @@ func (s *Scanner) recordBlobMetadata(
 }
 
 // cleanupBlobTempFile closes and removes the blob's temporary file
-func (s *Scanner) cleanupBlobTempFile(blobWithReader *blob.WithReader) {
+func (s *Scanner) cleanupBlobTempFile(blobWithReader *blob.WithReader) error {
 	if blobWithReader.TempFile != nil {
 		tempName := blobWithReader.TempFile.Name()
 
 		err := blobWithReader.TempFile.Close()
 		if err != nil {
-			log.Fatal("Failed to close temp file", "file", tempName, "error", err)
+			return fmt.Errorf("closing blob temp file: %w", err)
 		}
 
 		err = s.fs.Remove(tempName)
 		if err != nil {
-			log.Fatal("Failed to remove temp file", "file", tempName, "error", err)
+			return fmt.Errorf("removing blob temp file: %w", err)
 		}
 	}
+
+	return nil
 }
 
 // streamingChunkInfo tracks chunk metadata collected during streaming
@@ -1869,12 +1872,11 @@ func (s *Scanner) processFileStreaming(
 	return nil
 }
 
-// updateChunkStats counts a chunk that was not already stored, and its
-// size. The scan result's file counts, BytesScanned and BytesSkipped are
-// not touched here: the scan phase counts each file once.
+// updateChunkStats counts a chunk that was not already stored. The scan
+// result's file counts, BytesScanned and BytesSkipped are not touched
+// here: the scan phase counts each file once.
 func (s *Scanner) updateChunkStats(chunkSize int64, result *ScanResult) {
 	result.ChunksCreated++
-	result.NewChunkSize += chunkSize
 
 	if s.progress != nil {
 		s.progress.GetStats().ChunksCreated.Add(1)

internal/snapshot/snapshot.go

@@ -199,7 +199,6 @@ func (sm *SnapshotManager) UpdateSnapshotStatsExtended(
 			stats.CompressionLevel,
 			stats.BytesUploaded,
 			stats.UploadDurationMs,
-			stats.NewChunkSize,
 		)
 	})
 }
@@ -907,7 +906,6 @@ type ExtendedBackupStats struct {
 	BlobUncompressedSize int64 // Total uncompressed size of all referenced blobs
 	CompressionLevel     int   // Compression level used for this snapshot
 	UploadDurationMs     int64 // Total milliseconds spent uploading to S3
-	NewChunkSize         int64 // Total size of the chunks counted in ChunksCreated
 }
 
 // deleteOtherSnapshots deletes all snapshots except the current one

internal/vaultik/deep_verify_test.go

@@ -1,20 +1,15 @@
 package vaultik_test
 
 import (
-	"bytes"
 	"context"
-	"crypto/rand"
 	"io"
 	"os"
 	"path/filepath"
-	"slices"
 	"testing"
 
 	"github.com/spf13/afero"
 	"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/ui"
 	"sneak.berlin/go/vaultik/internal/vaultik"
 )
@@ -82,161 +77,6 @@ func TestDeepVerifyAcceptsHealthyAndRejectsCorruptBlob(t *testing.T) {
 		"deep verify should fail on a corrupted blob")
 }
 
-// TestDeepVerifyJSONCountsBlobs checks the counts that deep verify
-// reports under --json: on a healthy snapshot, with a blob that cannot
-// be checked, with a damaged blob, with missing blobs, and with a blob
-// of the wrong size. Deep verify stops at the first damaged blob, so the
-// blobs read before it are the ones verified; when blobs are missing or
-// the wrong size it reads no blob contents, but counts every such blob.
-func TestDeepVerifyJSONCountsBlobs(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")
-	dbPath := filepath.Join(tempDir, "index.sqlite")
-
-	chunkSize := int64(32 * 1024)
-	maxBlobSize := int64(128 * 1024)
-
-	// Random data does not compress or deduplicate, so it fills several
-	// blobs.
-	data := make([]byte, maxBlobSize*4)
-	_, err := rand.Read(data)
-	require.NoError(t, err)
-	require.NoError(t, fs.MkdirAll(dataDir, 0o755))
-	require.NoError(t, afero.WriteFile(fs,
-		filepath.Join(dataDir, "data.bin"), data, 0o644))
-
-	ctx := context.Background()
-
-	cfg, storer, snapshotID := runFileStorageBackup(
-		ctx, t, fs, dataDir, storeDir, dbPath, chunkSize, maxBlobSize)
-
-	healthy, err := deepVerifyJSON(ctx, t, cfg, storer, snapshotID)
-	require.NoError(t, err)
-	require.Equal(t, "ok", healthy.Status)
-	require.GreaterOrEqual(t, healthy.BlobCount, 3)
-	require.Equal(t, healthy.BlobCount, healthy.Verified)
-	require.Zero(t, healthy.Missing)
-
-	// Deep verify reads blobs in hash order, which is the sorted order of
-	// their storage keys.
-	keys, err := storer.List(ctx, "blobs/")
-	require.NoError(t, err)
-	slices.Sort(keys)
-	require.Len(t, keys, healthy.BlobCount)
-
-	// A blob that is present but cannot be checked is not missing: the
-	// check stops with the storage error.
-	denied := &statDeniedStorer{Storer: storer, deniedKey: keys[0]}
-
-	got, err := deepVerifyJSON(ctx, t, cfg, denied, snapshotID)
-	require.Error(t, err)
-	require.Equal(t, "failed", got.Status)
-	require.Zero(t, got.Missing)
-	require.Contains(t, got.ErrorMessage, os.ErrPermission.Error())
-
-	// Damage the last blob read, keeping its length, so every blob before
-	// it is read and passes first.
-	lastPath := filepath.Join(storeDir, keys[len(keys)-1])
-	blob, err := afero.ReadFile(fs, lastPath)
-	require.NoError(t, err)
-
-	blob[len(blob)/2] ^= 0xff
-	require.NoError(t, afero.WriteFile(fs, lastPath, blob, 0o600))
-
-	got, err = deepVerifyJSON(ctx, t, cfg, storer, snapshotID)
-	require.Error(t, err)
-	require.Equal(t, "failed", got.Status)
-	require.Equal(t, healthy.BlobCount-1, got.Verified)
-	require.Zero(t, got.Missing)
-
-	// Delete the first blob and the damaged last one: both must be
-	// counted as missing, not only the first one found.
-	var deletedSize int64
-
-	for _, key := range []string{keys[0], keys[len(keys)-1]} {
-		info, statErr := storer.Stat(ctx, key)
-		require.NoError(t, statErr)
-
-		deletedSize += info.Size
-
-		require.NoError(t, storer.Delete(ctx, key))
-	}
-
-	got, err = deepVerifyJSON(ctx, t, cfg, storer, snapshotID)
-	require.Error(t, err)
-	require.Equal(t, "failed", got.Status)
-	require.Equal(t, 2, got.Missing)
-	require.Equal(t, deletedSize, got.MissingSize)
-	require.Zero(t, got.Verified)
-
-	// Make a blob between the two deleted ones longer: it is counted as
-	// the wrong size, and both missing blobs are still counted.
-	grownPath := filepath.Join(storeDir, keys[1])
-	blob, err = afero.ReadFile(fs, grownPath)
-	require.NoError(t, err)
-
-	blob = append(blob, []byte("extra")...)
-	require.NoError(t, afero.WriteFile(fs, grownPath, blob, 0o600))
-
-	got, err = deepVerifyJSON(ctx, t, cfg, storer, snapshotID)
-	require.Error(t, err)
-	require.Equal(t, "failed", got.Status)
-	require.Equal(t, 1, got.Mismatched)
-	require.Equal(t, 2, got.Missing)
-	require.Zero(t, got.Verified)
-}
-
-// statDeniedStorer passes every call through to the wrapped Storer,
-// except that Stat of deniedKey fails with a permission error, as it
-// does for a blob that is present but cannot be read.
-type statDeniedStorer struct {
-	storage.Storer
-
-	deniedKey string
-}
-
-func (s *statDeniedStorer) Stat(
-	ctx context.Context, key string,
-) (*storage.ObjectInfo, error) {
-	if key == s.deniedKey {
-		return nil, os.ErrPermission
-	}
-
-	return s.Storer.Stat(ctx, key)
-}
-
-// deepVerifyJSON runs deep verify with --json on snapshotID and returns
-// the decoded result and the error RunDeepVerify returned.
-func deepVerifyJSON(
-	ctx context.Context, t *testing.T,
-	cfg *config.Config, storer storage.Storer, snapshotID string,
-) (vaultik.VerifyResult, error) {
-	t.Helper()
-
-	var out bytes.Buffer
-
-	v := &vaultik.Vaultik{
-		Config:  cfg,
-		Storage: storer,
-		Fs:      afero.NewOsFs(),
-		Stdout:  &out,
-		Stderr:  io.Discard,
-		UI:      ui.NewWithColor(io.Discard, false),
-	}
-	v.SetContext(ctx)
-
-	err := v.RunDeepVerify(snapshotID,
-		&vaultik.VerifyOptions{Deep: true, JSON: true})
-
-	return decodeVerifyResult(t, out.Bytes()), err
-}
-
 // corruptOneBlob flips a middle byte of the first blob file found under
 // blobsDir, leaving the file length unchanged.
 func corruptOneBlob(t *testing.T, fs afero.Fs, blobsDir string) {

internal/vaultik/new_chunk_size_test.go

@@ -1,133 +0,0 @@
-package vaultik_test
-
-import (
-	"bytes"
-	"context"
-	"maps"
-	"path/filepath"
-	"testing"
-
-	"github.com/spf13/afero"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-	"sneak.berlin/go/vaultik/internal/chunker"
-	"sneak.berlin/go/vaultik/internal/database"
-	"sneak.berlin/go/vaultik/internal/log"
-	"sneak.berlin/go/vaultik/internal/storage"
-	"sneak.berlin/go/vaultik/internal/vaultik"
-)
-
-// TestNewChunkSizeOfChangedFileStaysFixed covers
-// https://git.eeqj.de/sneak/vaultik/issues/311. A backup whose only change
-// is a modified file lists the size of the chunks that change added as its
-// NEW CHUNK SIZE, and a later backup that changes the file again leaves
-// every earlier snapshot's figure as it was. The file's row in the local
-// index is shared by all three snapshots and rewritten by each backup.
-func TestNewChunkSizeOfChangedFileStaysFixed(t *testing.T) {
-	log.Initialize(log.Config{})
-	t.Parallel()
-
-	fs := afero.NewOsFs()
-	tempDir := t.TempDir()
-	dataDir := filepath.Join(tempDir, "src")
-	dbPath := filepath.Join(tempDir, "index.sqlite")
-	changedPath := filepath.Join(dataDir, "changed.bin")
-
-	ctx := context.Background()
-	files := writeFaultSourceTree(t, fs, dataDir)
-	cfg := changedFileConfig(dataDir, dbPath)
-
-	appendRandomBytes(t, fs, files, changedPath, appendedFileSize)
-
-	store, err := storage.NewFileStorer(filepath.Join(tempDir, "remote"))
-	require.NoError(t, err)
-
-	db, err := database.New(ctx, dbPath)
-	require.NoError(t, err)
-	t.Cleanup(func() { _ = db.Close() })
-
-	repos := database.NewRepositories(db)
-	v := newBackupVaultik(ctx, cfg, store, repos, db, fs)
-
-	require.NoError(t, backUp(v, "first"))
-
-	firstFiles := maps.Clone(files)
-	firstID := localSnapshotID(ctx, t, repos, "first")
-
-	appendRandomBytes(t, fs, files, changedPath, faultChunkSize)
-	require.NoError(t, backUp(v, "second"))
-
-	secondID := localSnapshotID(ctx, t, repos, "second")
-	secondAdded := sizeOfNewChunks(t, files, firstFiles)
-	require.Positive(t, secondAdded)
-	assert.Equal(t, secondAdded, listedNewChunkSizes(t, v)[secondID],
-		"the second snapshot must count the chunks the change added")
-
-	appendRandomBytes(t, fs, files, changedPath, faultChunkSize)
-	require.NoError(t, backUp(v, "third"))
-
-	sizes := listedNewChunkSizes(t, v)
-	assert.Equal(t, sizeOfNewChunks(t, firstFiles, nil), sizes[firstID],
-		"a later backup must not change the first snapshot's figure")
-	assert.Equal(t, secondAdded, sizes[secondID],
-		"a later backup must not change the second snapshot's figure")
-}
-
-// sizeOfNewChunks returns the total size of the distinct chunks of the
-// files in after that no file in before has, cut the way the backup cuts
-// them.
-func sizeOfNewChunks(t *testing.T, after, before map[string][]byte) int64 {
-	t.Helper()
-
-	known := make(map[string]bool)
-	for _, chunk := range chunksOf(t, before) {
-		known[chunk.Hash] = true
-	}
-
-	var total int64
-
-	for _, chunk := range chunksOf(t, after) {
-		if !known[chunk.Hash] {
-			known[chunk.Hash] = true
-			total += chunk.Size
-		}
-	}
-
-	return total
-}
-
-// chunksOf returns the chunks of every file in files.
-func chunksOf(t *testing.T, files map[string][]byte) []chunker.Chunk {
-	t.Helper()
-
-	c := chunker.NewChunker(faultChunkSize)
-
-	var all []chunker.Chunk
-
-	for _, content := range files {
-		chunks, err := c.ChunkReader(bytes.NewReader(content))
-		require.NoError(t, err)
-
-		all = append(all, chunks...)
-	}
-
-	return all
-}
-
-// listedNewChunkSizes runs `snapshot list --json` and returns each listed
-// snapshot's new_chunk_size by snapshot ID.
-func listedNewChunkSizes(t *testing.T, v *vaultik.Vaultik) map[string]int64 {
-	t.Helper()
-
-	var out bytes.Buffer
-
-	v.Stdout = &out
-	require.NoError(t, v.ListSnapshots(true))
-
-	sizes := make(map[string]int64)
-	for _, row := range decodeListJSON(t, out.String()) {
-		sizes[row.ID] = row.NewChunkSize
-	}
-
-	return sizes
-}

internal/vaultik/restore.go

@@ -309,22 +309,11 @@ func (v *Vaultik) restoreAllFiles(
 		return nil, err
 	}
 
-	plan, unplanned, err := newRestorePlan(
-		v.ctx, repos, files, chunkToBlobMap, blobIDToHash)
+	plan, err := newRestorePlan(v.ctx, repos, files, chunkToBlobMap, blobIDToHash)
 	if err != nil {
 		return nil, fmt.Errorf("building restore plan: %w", err)
 	}
 
-	// Ranging over files rather than the map reports them in a stable order.
-	for _, file := range files {
-		if planErr, ok := unplanned[file.ID]; ok {
-			err = v.handleRestoreFileError(plan, opts, result, file, file.ID, planErr)
-			if err != nil {
-				return nil, err
-			}
-		}
-	}
-
 	filesByID, totalBytesExpected := indexRestoreFiles(files)
 
 	v.UI.Beginf("Restoring %s files (%s) to %s.",

internal/vaultik/restore_malformed_db_test.go

@@ -61,8 +61,8 @@ func craftedFile(
 }
 
 // TestRestoreShortChunkHashInNoBlob proves a file whose short chunk hash
-// has no blob_chunks row is left out of the restore plan with an error,
-// and fails the chunk write with an error.
+// has no blob_chunks row fails restore planning and the chunk write with
+// an error.
 func TestRestoreShortChunkHashInNoBlob(t *testing.T) {
 	t.Parallel()
 
@@ -77,11 +77,9 @@ func TestRestoreShortChunkHashInNoBlob(t *testing.T) {
 	chunkToBlobMap, err := v.buildChunkToBlobMap(ctx, repos)
 	require.NoError(t, err)
 
-	plan, unplanned, err := newRestorePlan(ctx, repos, []*database.File{file},
+	_, err = newRestorePlan(ctx, repos, []*database.File{file},
 		chunkToBlobMap, map[string]string{})
-	require.NoError(t, err)
-	require.ErrorIs(t, unplanned[file.ID], errPlanChunkMissing)
-	require.False(t, plan.hasPending())
+	require.ErrorIs(t, err, errPlanChunkMissing)
 
 	fileChunks, err := repos.FileChunks.GetByFileID(ctx, file.ID)
 	require.NoError(t, err)

internal/vaultik/restore_plan.go

@@ -41,25 +41,20 @@ type restorePlan struct {
 
 // newRestorePlan builds the file→blob index for the given files. Files
 // whose chunks reference no blobs (symlinks, directories) start in the
-// ready queue immediately. A file with a chunk that no blob holds cannot
-// be restored: it is left out of the plan and returned, by ID, with the
-// reason.
+// ready queue immediately.
 func newRestorePlan(
 	ctx context.Context,
 	repos *database.Repositories,
 	files []*database.File,
 	chunkToBlobMap map[string]*database.BlobChunk,
 	blobIDToHash map[string]string,
-) (*restorePlan, map[types.FileID]error, error) {
+) (*restorePlan, error) {
 	p := &restorePlan{
 		fileBlobs: make(map[types.FileID]map[string]struct{}, len(files)),
 		blobFiles: make(map[string]map[types.FileID]struct{}),
 		ready:     make([]types.FileID, 0, len(files)),
 		cached:    make(map[string]struct{}),
 	}
-	unplanned := make(map[types.FileID]error)
-
-nextFile:
 	for _, f := range files {
 		if f.IsSymlink() || f.Mode&uint32(os.ModeDir) != 0 {
 			// No chunks to fetch — restore can run immediately.
@@ -71,7 +66,7 @@ nextFile:
 
 		fileChunks, err := repos.FileChunks.GetByFileID(ctx, f.ID)
 		if err != nil {
-			return nil, nil, fmt.Errorf("planning %s: %w", f.Path, err)
+			return nil, fmt.Errorf("planning %s: %w", f.Path, err)
 		}
 
 		blobs := make(map[string]struct{})
@@ -79,18 +74,14 @@ nextFile:
 		for _, fc := range fileChunks {
 			bc, ok := chunkToBlobMap[fc.ChunkHash.String()]
 			if !ok {
-				unplanned[f.ID] = fmt.Errorf("%w: %s",
-					errPlanChunkMissing, shortHash(fc.ChunkHash.String()))
-
-				continue nextFile
+				return nil, fmt.Errorf("planning %s: %w: %s",
+					f.Path, errPlanChunkMissing, shortHash(fc.ChunkHash.String()))
 			}
 
 			hash, ok := blobIDToHash[bc.BlobID.String()]
 			if !ok {
-				unplanned[f.ID] = fmt.Errorf("%w: %s",
-					errPlanBlobIDMissing, bc.BlobID)
-
-				continue nextFile
+				return nil, fmt.Errorf("planning %s: %w: %s",
+					f.Path, errPlanBlobIDMissing, bc.BlobID)
 			}
 
 			blobs[hash] = struct{}{}
@@ -112,7 +103,7 @@ nextFile:
 		}
 	}
 
-	return p, unplanned, nil
+	return p, nil
 }
 
 // markBlobCached records that the named blob is now resident in the

internal/vaultik/restore_skip_errors_test.go

@@ -28,8 +28,7 @@ const (
 )
 
 // missingBlobBackup is a snapshot of single-chunk files spread over two
-// blobs, with one of those blobs lost: deleted from the store, or missing
-// from the snapshot metadata.
+// blobs, with one of those blobs deleted from the store.
 type missingBlobBackup struct {
 	fs         afero.Fs
 	cfg        *config.Config
@@ -38,39 +37,20 @@ type missingBlobBackup struct {
 	restoreDir string
 	// files holds the original content by source path.
 	files map[string][]byte
-	// lost holds the source paths whose chunk is in the lost blob.
+	// lost holds the source paths whose chunk is in the deleted blob.
 	lost map[string]bool
 }
 
 // TestRestoreSkipErrorsSkipsFilesOfMissingBlob restores with SkipErrors
-// after one blob of a two-blob snapshot was deleted from the store.
+// after one blob of a two-blob snapshot was deleted. Every file stored in
+// that blob must be reported as failed and left absent, every other file
+// must be restored intact, and Restore must still return an error.
 func TestRestoreSkipErrorsSkipsFilesOfMissingBlob(t *testing.T) {
 	log.Initialize(log.Config{})
 	t.Parallel()
 
 	ctx := context.Background()
-	assertRestoreSkipsLostFiles(ctx, t, backupThenDeleteOneBlob(ctx, t))
-}
-
-// TestRestoreSkipErrorsSkipsFilesWithChunkInNoBlob restores with
-// SkipErrors a snapshot whose metadata lists chunks that no blob holds, as
-// earlier backup defects left behind.
-func TestRestoreSkipErrorsSkipsFilesWithChunkInNoBlob(t *testing.T) {
-	log.Initialize(log.Config{})
-	t.Parallel()
-
-	ctx := context.Background()
-	assertRestoreSkipsLostFiles(ctx, t, backupThenDropOneBlobFromMetadata(ctx, t))
-}
-
-// assertRestoreSkipsLostFiles restores backup with SkipErrors. Every file
-// stored in the lost blob must be reported as failed and left absent,
-// every other file must be restored intact, and Restore must still return
-// an error.
-func assertRestoreSkipsLostFiles(
-	ctx context.Context, t *testing.T, backup *missingBlobBackup,
-) {
-	t.Helper()
+	backup := backupThenDeleteOneBlob(ctx, t)
 
 	var out bytes.Buffer
 
@@ -92,14 +72,14 @@ func assertRestoreSkipsLostFiles(
 
 		if backup.lost[path] {
 			assert.Containsf(t, out.String(), path,
-				"%s needs the lost blob and must be reported", path)
+				"%s needs the deleted blob and must be reported", path)
 			assert.NoFileExists(t, restored)
 
 			continue
 		}
 
 		got, err := afero.ReadFile(backup.fs, restored)
-		require.NoErrorf(t, err, "%s does not need the lost blob", path)
+		require.NoErrorf(t, err, "%s does not need the deleted blob", path)
 		assert.Equalf(t, content, got, "%s restored with wrong content", path)
 	}
 }
@@ -123,30 +103,6 @@ func TestRestoreMissingBlobAbortsWithoutSkipErrors(t *testing.T) {
 	require.ErrorIs(t, err, storage.ErrNotFound)
 }
 
-// TestRestoreChunkInNoBlobAbortsWithoutSkipErrors checks that a chunk held
-// by no blob still ends the restore, before any file is written, when
-// SkipErrors is not set.
-func TestRestoreChunkInNoBlobAbortsWithoutSkipErrors(t *testing.T) {
-	log.Initialize(log.Config{})
-	t.Parallel()
-
-	ctx := context.Background()
-	backup := backupThenDropOneBlobFromMetadata(ctx, t)
-
-	v := newReaderVaultik(ctx, backup.cfg, backup.storer, nil, backup.fs)
-
-	err := v.Restore(&vaultik.RestoreOptions{
-		SnapshotID: backup.snapshotID,
-		TargetDir:  backup.restoreDir,
-	})
-
-	require.ErrorContains(t, err, "chunk missing from blob map")
-
-	for path := range backup.files {
-		assert.NoFileExists(t, filepath.Join(backup.restoreDir, path))
-	}
-}
-
 // backupThenDeleteOneBlob backs up missingBlobFileCount single-chunk
 // files, then deletes from the store the blob holding the first of them.
 func backupThenDeleteOneBlob(
@@ -154,54 +110,6 @@ func backupThenDeleteOneBlob(
 ) *missingBlobBackup {
 	t.Helper()
 
-	backup, db, lostBlob := backupSingleChunkFiles(ctx, t)
-	require.NoError(t, db.Close())
-
-	require.NoError(t, backup.storer.Delete(ctx, fmt.Sprintf(
-		"blobs/%s/%s/%s", lostBlob[:2], lostBlob[2:4], lostBlob)))
-
-	return backup
-}
-
-// backupThenDropOneBlobFromMetadata backs up missingBlobFileCount
-// single-chunk files, then removes the blob holding the first of them from
-// the snapshot and exports the snapshot metadata again. The exported
-// metadata then lists chunks that no blob holds.
-func backupThenDropOneBlobFromMetadata(
-	ctx context.Context, t *testing.T,
-) *missingBlobBackup {
-	t.Helper()
-
-	backup, db, lostBlob := backupSingleChunkFiles(ctx, t)
-
-	// The uploads row refers to the blob row, so it would keep the export
-	// from deleting it.
-	for _, query := range []string{
-		"DELETE FROM snapshot_blobs WHERE blob_hash = ?",
-		"DELETE FROM uploads WHERE blob_hash = ?",
-	} {
-		_, err := db.Conn().ExecContext(ctx, query, lostBlob)
-		require.NoError(t, err)
-	}
-
-	sm := newFaultSnapshotManager(
-		backup.fs, backup.storer, backup.cfg, database.NewRepositories(db))
-	require.NoError(t,
-		sm.ExportSnapshotMetadata(ctx, db.Path(), backup.snapshotID))
-	require.NoError(t, db.Close())
-
-	return backup
-}
-
-// backupSingleChunkFiles backs up missingBlobFileCount single-chunk files.
-// It returns the backup, with lost naming the files stored in the same
-// blob as the first of them, the still open local index, and the hash of
-// that blob.
-func backupSingleChunkFiles(
-	ctx context.Context, t *testing.T,
-) (*missingBlobBackup, *database.DB, string) {
-	t.Helper()
-
 	fs := afero.NewOsFs()
 	tempDir := t.TempDir()
 	dataDir := filepath.Join(tempDir, "src")
@@ -235,11 +143,16 @@ func backupSingleChunkFiles(
 		blobOfFile[path] = blobHashOfSingleChunkFile(ctx, t, repos, path)
 	}
 
-	lostBlob := blobOfFile[filepath.Join(dataDir, "file-00.bin")]
+	require.NoError(t, db.Close())
+
+	deleted := blobOfFile[filepath.Join(dataDir, "file-00.bin")]
+	require.NoError(t, storer.Delete(ctx, fmt.Sprintf(
+		"blobs/%s/%s/%s", deleted[:2], deleted[2:4], deleted)))
+
 	lost := make(map[string]bool)
 
 	for path, hash := range blobOfFile {
-		if hash == lostBlob {
+		if hash == deleted {
 			lost[path] = true
 		}
 	}
@@ -255,7 +168,7 @@ func backupSingleChunkFiles(
 		restoreDir: filepath.Join(tempDir, "restored"),
 		files:      files,
 		lost:       lost,
-	}, db, lostBlob
+	}
 }
 
 // blobHashOfSingleChunkFile returns the hash of the blob holding the one

internal/vaultik/snapshot.go

@@ -209,7 +209,6 @@ type snapshotStats struct {
 	totalFiles         int
 	totalBytes         int64
 	totalChunks        int
-	newChunkSize       int64
 	totalBlobs         int
 	totalBytesSkipped  int64
 	totalFilesSkipped  int
@@ -342,7 +341,6 @@ func (v *Vaultik) scanAllDirectories(
 		stats.totalFiles += result.FilesScanned
 		stats.totalBytes += result.BytesScanned
 		stats.totalChunks += result.ChunksCreated
-		stats.newChunkSize += result.NewChunkSize
 		stats.totalBlobs += result.BlobsCreated
 		stats.totalFilesSkipped += result.FilesSkipped
 		stats.totalFilesFailed += result.FilesFailed
@@ -405,7 +403,6 @@ func (v *Vaultik) finalizeSnapshotMetadata(
 		BlobUncompressedSize: stats.blobUncompressedSize,
 		CompressionLevel:     v.Config.CompressionLevel,
 		UploadDurationMs:     stats.uploadDuration.Milliseconds(),
-		NewChunkSize:         stats.newChunkSize,
 	}
 
 	err = v.SnapshotManager.UpdateSnapshotStatsExtended(v.ctx, snapshotID, extStats)

internal/vaultik/snapshot_list_test.go

@@ -487,7 +487,6 @@ type listJSONRow struct {
 	RemoteKey      string `json:"remote_key"`
 	Timestamp      string `json:"timestamp"`
 	CompressedSize int64  `json:"compressed_size"`
-	NewChunkSize   int64  `json:"new_chunk_size"`
 	LocallyTracked bool   `json:"locally_tracked"`
 	RemotePresent  *bool  `json:"remote_present"`
 }

internal/vaultik/verify.go

@@ -17,7 +17,6 @@ import (
 	"sneak.berlin/go/vaultik/internal/database"
 	"sneak.berlin/go/vaultik/internal/log"
 	"sneak.berlin/go/vaultik/internal/snapshot"
-	"sneak.berlin/go/vaultik/internal/storage"
 )
 
 // Sentinel errors for snapshot verification failures.
@@ -33,9 +32,7 @@ var (
 	errManifestExtraBlob   = errors.New("manifest contains blob not in database")
 	errManifestMissingBlob = errors.New(
 		"manifest omits blob present in database")
-	errBlobSizeMismatch  = errors.New("blob size mismatch")
-	errBlobsNotInStorage = errors.New(
-		"blobs missing from storage or stored with the wrong size")
+	errBlobSizeMismatch = errors.New("blob size mismatch")
 )
 
 // verifyStatusFailed is the JSON status value for a failed verification.
@@ -138,6 +135,7 @@ func (v *Vaultik) RunDeepVerify(snapshotID string, opts *VerifyOptions) error {
 	}
 
 	result.Status = "ok"
+	result.Verified = len(dbBlobs)
 
 	if opts.JSON {
 		return v.outputVerifyJSON(result)
@@ -292,7 +290,7 @@ func (v *Vaultik) runVerificationSteps(
 		v.stdoutf("Checking blob existence in remote storage...\n")
 	}
 
-	err = v.verifyBlobExistenceFromDB(dbBlobs, result)
+	err = v.verifyBlobExistenceFromDB(dbBlobs)
 	if err != nil {
 		return v.deepVerifyFailure(result, opts, err.Error(), err)
 	}
@@ -303,8 +301,7 @@ func (v *Vaultik) runVerificationSteps(
 			len(dbBlobs), ubytes(totalSize))
 	}
 
-	err = v.performDeepVerificationFromDB(
-		dbBlobs, tdb.db.Conn(), opts, result, identities)
+	err = v.performDeepVerificationFromDB(dbBlobs, tdb.db.Conn(), opts, identities)
 	if err != nil {
 		return v.deepVerifyFailure(result, opts, err.Error(), err)
 	}
@@ -656,14 +653,8 @@ func (v *Vaultik) verifyManifestAgainstDatabase(
 	return nil
 }
 
-// verifyBlobExistenceFromDB checks that every blob from the database is
-// in storage with the size the database records. It checks them all,
-// counting the missing and wrong-size blobs in result, and fails if it
-// found any. Any storage error other than not found stops the check and
-// fails with that error.
-func (v *Vaultik) verifyBlobExistenceFromDB(
-	blobs []snapshot.BlobInfo, result *VerifyResult,
-) error {
+// verifyBlobExistenceFromDB checks that all blobs from database exist in S3
+func (v *Vaultik) verifyBlobExistenceFromDB(blobs []snapshot.BlobInfo) error {
 	log.Info("Verifying blob existence in S3", "blob_count", len(blobs))
 
 	for i, blob := range blobs {
@@ -676,23 +667,17 @@ func (v *Vaultik) verifyBlobExistenceFromDB(
 		// Construct blob path
 		blobPath := fmt.Sprintf("blobs/%s/%s/%s", blob.Hash[:2], blob.Hash[2:4], blob.Hash)
 
+		// Check blob exists
 		stat, err := v.Storage.Stat(v.ctx, blobPath)
-		switch {
-		case errors.Is(err, storage.ErrNotFound):
-			log.Warn("Blob missing from storage", "hash", blob.Hash)
+		if err != nil {
+			return fmt.Errorf("blob %s missing from storage: %w", blob.Hash, err)
+		}
 
-			result.Missing++
-			result.MissingSize += blob.CompressedSize
-		case err != nil:
-			return fmt.Errorf("checking blob %s in storage: %w",
-				blob.Hash, err)
-		case stat.Size != blob.CompressedSize:
-			log.Warn("Blob has the wrong size in storage",
-				"hash", blob.Hash,
-				"storage_size", stat.Size,
-				"database_size", blob.CompressedSize)
-
-			result.Mismatched++
+		// Verify size matches
+		if stat.Size != blob.CompressedSize {
+			return fmt.Errorf(
+				"%w: blob %s: S3 has %d bytes, database has %d bytes",
+				errBlobSizeMismatch, blob.Hash, stat.Size, blob.CompressedSize)
 		}
 
 		// Progress update every 100 blobs
@@ -706,22 +691,16 @@ func (v *Vaultik) verifyBlobExistenceFromDB(
 		}
 	}
 
-	if result.Missing > 0 || result.Mismatched > 0 {
-		return fmt.Errorf("%w: %d missing, %d with the wrong size",
-			errBlobsNotInStorage, result.Missing, result.Mismatched)
-	}
-
 	log.Info("✓ All blobs exist in storage")
 
 	return nil
 }
 
 // performDeepVerificationFromDB downloads and verifies the content of
-// each blob using the database as source, stopping at the first blob that
-// fails. Each blob that passes is counted in result.Verified.
+// each blob using the database as source.
 func (v *Vaultik) performDeepVerificationFromDB(
 	blobs []snapshot.BlobInfo, db *sql.DB, opts *VerifyOptions,
-	result *VerifyResult, identities []age.Identity,
+	identities []age.Identity,
 ) error {
 	// Calculate total bytes for ETA
 	var totalBytesExpected int64
@@ -744,7 +723,6 @@ func (v *Vaultik) performDeepVerificationFromDB(
 			return fmt.Errorf("blob %s verification failed: %w", blobInfo.Hash, err)
 		}
 
-		result.Verified++
 		bytesProcessed += blobInfo.CompressedSize
 		elapsed := time.Since(startTime)
 		remaining := len(blobs) - (i + 1)