What causes "get object by doc id: %w" in weaviate/weaviate?

Inspect the wrapped inner error for the underlying store failure

get object by doc id: %w

Error message

get object by doc id: %w

What it means

uuidFromDocID resolves a doc ID to a UUID by looking up the object in the bucket via its secondary index (doc ID key). The GetBySecondary call failed, so the reverse lookup cannot proceed. Used by FindUUIDs (e.g. during geo/filtered scans and reference resolution).

Source

Thrown at adapters/repos/db/shard_read.go:1011

	if err != nil {
		return nil, errors.Wrap(err, "build inverted filter allow list")
	}

	return list, nil
}

func (s *Shard) uuidFromDocID(docID uint64) (strfmt.UUID, error) {
	bucket, release, err := s.objectsBucket()
	if err != nil {
		return "", err
	}
	defer release()

	docIDBytes := make([]byte, 8)
	binary.LittleEndian.PutUint64(docIDBytes, docID)
	res, err := bucket.GetBySecondary(context.TODO(), 0, docIDBytes) // TODO: context
	if err != nil {
		return "", fmt.Errorf("get object by doc id: %w", err)
	}

	prop, _, err := storobj.ParseAndExtractProperty(res, "id")
	if err != nil {
		return "", fmt.Errorf("parse and extract property: %w", err)
	}

	return strfmt.UUID(prop[0]), nil
}

func (s *Shard) batchDeleteObject(ctx context.Context, id strfmt.UUID, deletionTime time.Time) error {
	// Wait outside RLock: initAsyncReplication holds the write lock while
	// initialising, so blocking under RLock here would deadlock.
	if err := s.waitForMinimalHashTreeInitialization(ctx); err != nil {
		return err
	}

	s.asyncReplicationRWMux.RLock()

View on GitHub (pinned to 75aa4b6d11)

Solutions

Inspect the wrapped inner error for the underlying store failure
Retry the query — transient errors occur if the shard is being dropped or migrated concurrently
If persistent, verify shard integrity and restore from backup or rebuild
Check disk health and node logs for LSM store issues

Defensive patterns

Strategy: retry

Validate before calling

// Ensure the target collection/shard is READY before doc-ID-dependent queries
status, _ := client.Cluster().NodesStatus(ctx)

Try / catch

uuids, err := shard.FindUUIDs(ctx, docIDs)
if err != nil && strings.Contains(err.Error(), "get object by doc id") {
	// transient if shard was dropping/migrating — retry with backoff
}

Prevention

Retry idempotent read paths with backoff around shard lifecycle events
Monitor disk health and LSM store warnings in node logs
Prevent concurrent shard deletion and heavy reads where possible
Restore corrupted shards from backup instead of continuing reads

When it happens

Trigger: FindUUIDs iterating doc IDs where bucket.GetBySecondary(0, docIDBytes) fails — LSM store read error, bucket closed (shard dropping), or IO failure during the lookup.

Common situations: Shard being dropped/migrated concurrently while a query resolves doc IDs; disk IO errors; corruption of the secondary index (docID→UUID mapping) after an unclean shutdown.

Related errors

AI-assisted analysis of weaviate/weaviate@75aa4b6d11 (2026-09-04). Data as JSON: /api/errors/fb4349c4a253a5aa. Report an issue: GitHub.