dgraph-io/dgraph · error

br.Close

Error message

br.Close

What it means

After mapping, RunMapper explicitly closes the backup reader; if br.Close() returns an error (e.g. failure flushing or closing the underlying object-store stream), it is wrapped as 'br.Close' and aborts the restore. Note there is also a deferred br.Close(); this is the checked, error-propagating close.

Source

Thrown at worker/restore_map.go:858

			in := &loadBackupInput{
				preds:     predSet,
				dropNs:    localDropNs,
				version:   manifest.Version,
				restoreTs: req.RestoreTs,
				// Only map the schema keys corresponding to the latest backup.
				keepSchema:              i == 0,
				compression:             manifest.Compression,
				fromNamespace:           req.FromNamespace,
				isNamespaceAwareRestore: req.IsNamespaceAwareRestore,
			}

			// This would stream the backups from the source, and map them in
			// Dgraph compatible format on disk.
			if err := mapper.Map(br, in); err != nil {
				return nil, errors.Wrap(err, "mapper.Map")
			}
			if err := br.Close(); err != nil {
				return nil, errors.Wrap(err, "br.Close")
			}
		}
		for _, op := range manifest.DropOperations {
			switch op.DropOp {
			case pb.DropOperation_ALL:
				dropAll = true
			case pb.DropOperation_DATA:
				if op.DropValue == "" {
					// In 2103, we do not support namespace level drop data.
					dropAll = true
					continue
				}
				ns, err := strconv.ParseUint(op.DropValue, 0, 64)
				if err != nil {
					return nil, errors.Wrap(err, "map phase failed to parse namespace")
				}
				dropNs[ns] = struct{}{}
			case pb.DropOperation_ATTR:

View on GitHub (pinned to 759e242be6)

Solutions

1. Read the wrapped cause; if it is a network/stream error, simply retry the restore.
2. Check object-store client timeout/retry configuration and network stability between the node and the storage backend.
3. For file:// locations, check the filesystem/disk health of the backup path.
4. If the error reproduces at the same point, the backup object may be corrupt — verify or re-take the backup.

// check network path to the object store before starting
if err := probeBackupLocation(req.Location, creds); err != nil { return err }

if _, err := worker.RunMapper(req, mapDir); err != nil {
    if strings.Contains(err.Error(), "br.Close") && isTransient(errors.Cause(err)) {
        return retryRestore(req, mapDir) // safe: restore aborted before commit
    }
    return err
}

When it happens

Trigger: br.Close() errors while shutting down the streamed backup object — typically an underlying network/S3 read stream failure, checksum verification on close, or descriptor close error.

Common situations: Object-store connection dropped mid/close during long restores; TLS timeouts to S3/GCS/Azure; local file reader for file:// backups hitting I/O errors.

Related errors

• cannot retrieve manifests
• while retrieving manifests
• newBackupReader
• unable to reach quorum
• unique validation failed to fetch schema for predicates %v

AI-assisted analysis of dgraph-io/dgraph@759e242be6 (2026-09-01). Data as JSON: /api/errors/ca259f8e8605f283. Report an issue: GitHub.