sneak/vaultik
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Populate snapshot_blobs for dedup-referenced blobs at completion

Browse Source 
The bug: fully-deduplicated snapshots (every chunk already in storage
from a prior run) had an empty snapshot_blobs table. The metadata-
export pipeline then dropped all blob/blob_chunks rows from the
exported database, leaving file_chunks references to chunks whose
blobs were no longer recorded. Restore fails on every file with
"chunk X not found in any blob".

Fix: at CompleteSnapshot time, run an INSERT OR IGNORE that links
every blob holding a chunk referenced by this snapshot's files into
snapshot_blobs. New blobs uploaded during the snapshot are already
recorded (no-op for them); dedup-referenced blobs are added.

The cleanup query in deleteOrphanedBlobs already restricts to
snapshot_blobs entries for the current snapshot — so once
snapshot_blobs is correctly populated, the exported database
contains the full set of blob/blob_chunks rows needed for restore.

Regression test: TestDedupOnlySnapshotRestores creates two
identical snapshots (the second uploads zero new blobs) and
restores the second. Without the fix, restore fails on every file.
This commit is contained in:
Jeffrey Paul
2026-06-17 06:05:52 +02:00
parent f1ce085972
commit 132f7149ca
3 changed files with 166 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
internal/database/snapshots.go
View File

@@ -331,6 +331,43 @@ func (r *SnapshotRepository) AddFilesByIDBatch(ctx context.Context, tx *sql.Tx,
return nil
}
// PopulateReferencedBlobs ensures snapshot_blobs contains an entry for
// every blob that holds a chunk referenced by any file in the snapshot.
// This is necessary because the AddBlob hook only runs when a blob is
// newly uploaded during a snapshot — fully-deduplicated snapshots (where
// every chunk already exists in storage from a prior run) would otherwise
// have an empty snapshot_blobs set and be impossible to restore.
//
// Returns the number of rows inserted (i.e. blobs that were previously
// referenced indirectly via file_chunks but not yet recorded in
// snapshot_blobs for this snapshot).
func (r *SnapshotRepository) PopulateReferencedBlobs(ctx context.Context, tx *sql.Tx, snapshotID string) (int64, error) {
query := `
INSERT OR IGNORE INTO snapshot_blobs (snapshot_id, blob_id, blob_hash)
SELECT DISTINCT ?, blobs.id, blobs.blob_hash
FROM blobs
JOIN blob_chunks ON blob_chunks.blob_id = blobs.id
JOIN file_chunks ON file_chunks.chunk_hash = blob_chunks.chunk_hash
JOIN snapshot_files ON snapshot_files.file_id = file_chunks.file_id
WHERE snapshot_files.snapshot_id = ?
AND blobs.blob_hash IS NOT NULL
`
var result sql.Result
var err error
if tx != nil {
result, err = tx.ExecContext(ctx, query, snapshotID, snapshotID)
} else {
result, err = r.db.ExecWithLog(ctx, query, snapshotID, snapshotID)
}
if err != nil {
return 0, fmt.Errorf("populating referenced blobs: %w", err)
}
n, _ := result.RowsAffected()
return n, nil
}
// AddBlob adds a blob to a snapshot
func (r *SnapshotRepository) AddBlob(ctx context.Context, tx *sql.Tx, snapshotID string, blobID types.BlobID, blobHash types.BlobHash) error {
query := `