benbjohnson/litestream · error

read txid file for crash recovery

Error message

read txid file for crash recovery: %w

What it means

During follow-mode crash recovery, ReadTXIDFile failed to read the <db>-txid sidecar file even though the database file exists. The sidecar is missing, unreadable, or not valid JSON, so the last applied TXID cannot be recovered.

Solutions

1. Restore permissions on the -txid sidecar file
2. If the sidecar was lost, delete the database (and sidecar) and re-restore from scratch

When it happens

Trigger: Thrown at replica.go:628 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of benbjohnson/litestream@4ed7a308f6 (2026-09-06). Data as JSON: /api/errors/c5bca52f91bf394a. Report an issue: GitHub.

Appendix: source

Thrown at replica.go:628

	if opt.OutputPath == "" {
		return fmt.Errorf("output path required")
	} else if opt.TXID != 0 && !opt.Timestamp.IsZero() {
		return fmt.Errorf("cannot specify index & timestamp to restore")
	} else if opt.Follow && opt.TXID != 0 {
		return fmt.Errorf("cannot use follow mode with -txid")
	} else if opt.Follow && !opt.Timestamp.IsZero() {
		return fmt.Errorf("cannot use follow mode with -timestamp")
	} else if opt.IntegrityCheck != IntegrityCheckNone && opt.IntegrityCheck != IntegrityCheckQuick && opt.IntegrityCheck != IntegrityCheckFull {
		return fmt.Errorf("unsupported integrity check mode: %d", opt.IntegrityCheck)
	}

	// In follow mode, if the database already exists, attempt crash recovery
	// by reading the last applied TXID from the sidecar file.
	if opt.Follow {
		if _, statErr := os.Stat(opt.OutputPath); statErr == nil {
			txid, readErr := ReadTXIDFile(opt.OutputPath)
			if readErr != nil {
				return fmt.Errorf("read txid file for crash recovery: %w", readErr)
			}
			if txid == 0 {
				return fmt.Errorf("cannot resume follow mode: database exists but no -txid file found; delete the database to re-restore: %s", opt.OutputPath)
			}
			// Validate saved TXID is still reachable. If the earliest snapshot
			// starts after our saved TXID, retention has pruned the history
			// and we can't catch up incrementally.
			snapshotItr, itrErr := r.Client.LTXFiles(ctx, SnapshotLevel, 0, false)
			if itrErr != nil {
				return fmt.Errorf("cannot validate saved TXID for crash recovery: %w", itrErr)
			}

			var latestSnapshot *ltx.FileInfo
			for snapshotItr.Next() {
				latestSnapshot = snapshotItr.Item()
			}
			if err := snapshotItr.Err(); err != nil {
				_ = snapshotItr.Close()

View on GitHub (pinned to 4ed7a308f6)