Record each snapshot's uncompressed size when its backup runs (closes #360)

snapshot list summed the chunks a snapshot reaches through its file rows
in the local index. A file's row is shared by every snapshot that lists
it and rewritten when the file changes, so backing up a changed file
changed every older snapshot's UNCOMPRESSED SIZE.

The backup now takes that total when it records the snapshot's stats,
into a new snapshots column, uncompressed_chunk_size, and snapshot list
reads the column, as it already does for new_chunk_size.

Model: opus-5-5

README.md

@@ -294,7 +294,8 @@ needed.
 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).
+were not stored before). Both chunk sizes are recorded when the
+snapshot's backup runs.
 
 Listing the destination store does **not** require the age secret key,
 so it works in vaultik's intended configuration, where the backed-up

docs/DATAMODEL.md

@@ -122,6 +122,7 @@ Tracks backup snapshots.
 - `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
+- `uncompressed_chunk_size` (INTEGER) - Total size before compression of the distinct chunks the snapshot's files reference, recorded when its backup ran; `snapshot list` shows it as UNCOMPRESSED SIZE
 - `blob_size` (INTEGER) - Total compressed size of all referenced blobs
 - `blob_uncompressed_size` (INTEGER) - Total uncompressed size of all referenced blobs
 - `compression_ratio` (REAL) - Compression ratio achieved

internal/database/schema/001.sql

@@ -89,6 +89,7 @@ CREATE TABLE IF NOT EXISTS snapshots (
     new_chunk_size INTEGER NOT NULL DEFAULT 0,
     blob_count INTEGER NOT NULL DEFAULT 0,
     total_size INTEGER NOT NULL DEFAULT 0,
+    uncompressed_chunk_size INTEGER NOT NULL DEFAULT 0,
     blob_size INTEGER NOT NULL DEFAULT 0,
     blob_uncompressed_size INTEGER NOT NULL DEFAULT 0,
     compression_ratio REAL NOT NULL DEFAULT 1.0,

internal/database/snapshots.go

@@ -120,7 +120,11 @@ func (r *SnapshotRepository) UpdateCounts(
 	return nil
 }
 
-// UpdateExtendedStats updates extended statistics for a snapshot
+// UpdateExtendedStats updates extended statistics for a snapshot. It also
+// records the total plaintext size of the distinct chunks the snapshot's
+// files reference. Only the snapshot's own backup can take that total: a
+// file's row is shared by every snapshot that lists it, and a later
+// backup of a changed file rewrites the row with the new chunks.
 func (r *SnapshotRepository) UpdateExtendedStats(
 	ctx context.Context,
 	tx *sql.Tx,
@@ -145,7 +149,17 @@ func (r *SnapshotRepository) UpdateExtendedStats(
 		    compression_level = ?,
 		    upload_bytes = ?,
 		    upload_duration_ms = ?,
-		    new_chunk_size = ?
+		    new_chunk_size = ?,
+		    uncompressed_chunk_size = (
+		        SELECT COALESCE(SUM(c.size), 0)
+		        FROM (
+		            SELECT DISTINCT fc.chunk_hash
+		            FROM snapshot_files sf
+		            JOIN file_chunks fc ON sf.file_id = fc.file_id
+		            WHERE sf.snapshot_id = snapshots.id
+		        ) sc
+		        JOIN chunks c ON sc.chunk_hash = c.chunk_hash
+		    )
 		WHERE id = ?
 	`
 
@@ -570,30 +584,22 @@ func (r *SnapshotRepository) GetSnapshotBlobSizes(
 	return compressed, uncompressed, nil
 }
 
-// GetSnapshotUncompressedChunkSize returns the sum of plaintext sizes of all unique
-// chunks referenced by a snapshot (via snapshot_files → file_chunks → chunks).
+// GetSnapshotUncompressedChunkSize returns the total plaintext size of the
+// distinct chunks the snapshot's files reference, as the snapshot's backup
+// recorded it (see UpdateExtendedStats).
 func (r *SnapshotRepository) GetSnapshotUncompressedChunkSize(
 	ctx context.Context, snapshotID string,
 ) (int64, error) {
-	query := `
-		SELECT COALESCE(SUM(c.size), 0)
-		FROM (
-			SELECT DISTINCT fc.chunk_hash
-			FROM snapshot_files sf
-			JOIN file_chunks fc ON sf.file_id = fc.file_id
-			WHERE sf.snapshot_id = ?
-		) sc
-		JOIN chunks c ON sc.chunk_hash = c.chunk_hash
-	`
+	query := `SELECT uncompressed_chunk_size FROM snapshots WHERE id = ?`
 
-	var totalSize int64
+	var uncompressedChunkSize int64
 
-	err := r.db.conn.QueryRowContext(ctx, query, snapshotID).Scan(&totalSize)
+	err := r.db.conn.QueryRowContext(ctx, query, snapshotID).Scan(&uncompressedChunkSize)
 	if err != nil {
 		return 0, fmt.Errorf("querying uncompressed chunk size: %w", err)
 	}
 
-	return totalSize, nil
+	return uncompressedChunkSize, nil
 }
 
 // GetSnapshotNewChunkSize returns the total plaintext size of the chunks the

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/snapshot_list_test.go

@@ -483,13 +483,14 @@ func TestListSnapshots_UnreadableManifestDoesNotHideOthers(t *testing.T) {
 //
 //nolint:tagliatelle // snake_case is the established output format
 type listJSONRow struct {
-	ID             string `json:"id"`
-	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"`
+	ID               string `json:"id"`
+	RemoteKey        string `json:"remote_key"`
+	Timestamp        string `json:"timestamp"`
+	CompressedSize   int64  `json:"compressed_size"`
+	UncompressedSize int64  `json:"uncompressed_size"`
+	NewChunkSize     int64  `json:"new_chunk_size"`
+	LocallyTracked   bool   `json:"locally_tracked"`
+	RemotePresent    *bool  `json:"remote_present"`
 }
 
 // decodeListJSON parses the command's stdout, which must contain

internal/vaultik/uncompressed_size_test.go

@@ -0,0 +1,85 @@
+package vaultik_test
+
+import (
+	"bytes"
+	"context"
+	"path/filepath"
+	"testing"
+
+	"github.com/spf13/afero"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"sneak.berlin/go/vaultik/internal/database"
+	"sneak.berlin/go/vaultik/internal/log"
+	"sneak.berlin/go/vaultik/internal/storage"
+	"sneak.berlin/go/vaultik/internal/vaultik"
+)
+
+// TestUncompressedSizeOfChangedFileStaysFixed covers
+// https://git.eeqj.de/sneak/vaultik/issues/360. A backup that changes a
+// file leaves the earlier snapshot's UNCOMPRESSED SIZE as it was when that
+// snapshot's backup completed. The file's row in the local index is shared
+// by both snapshots and rewritten by the second backup.
+func TestUncompressedSizeOfChangedFileStaysFixed(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"))
+
+	// With nothing to compare against, sizeOfNewChunks totals every
+	// distinct chunk of the files.
+	firstID := localSnapshotID(ctx, t, repos, "first")
+	firstSize := sizeOfNewChunks(t, files, nil)
+	require.Equal(t, firstSize, listedUncompressedSizes(t, v)[firstID],
+		"the first snapshot must count the chunks of every file it holds")
+
+	appendRandomBytes(t, fs, files, changedPath, faultChunkSize)
+	require.NoError(t, backUp(v, "second"))
+
+	secondID := localSnapshotID(ctx, t, repos, "second")
+	sizes := listedUncompressedSizes(t, v)
+	assert.Equal(t, firstSize, sizes[firstID],
+		"a later backup must not change the first snapshot's figure")
+	assert.Equal(t, sizeOfNewChunks(t, files, nil), sizes[secondID],
+		"the second snapshot must count the changed file's chunks")
+}
+
+// listedUncompressedSizes runs `snapshot list --json` and returns each
+// listed snapshot's uncompressed_size by snapshot ID.
+func listedUncompressedSizes(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.UncompressedSize
+	}
+
+	return sizes
+}

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)