rqlite/rqlite · error

load in progress

Error message

load in progress

What it means

ErrLoadInProgress is returned when a load (database restore) is already in progress and a snapshot cannot be taken. In Notify/CAS-protected snapshot paths the store inspects the Raft snapshot error and, if it indicates a load is in progress, surfaces ErrLoadInProgress (store.go:2995). Snapshots are unsafe during a restore because the database files are being replaced.

Source

Thrown at store/store.go:105

	ErrWaitForFSMTimeout = errors.New("timeout waiting for fsm")

	// ErrInvalidBackupFormat is returned when the requested backup format
	// is not valid.
	ErrInvalidBackupFormat = errors.New("invalid backup format")

	// ErrBackupCASFailed is returned when we cannot acquire the backup CAS lock.
	ErrBackupCASFailed = errors.New("failed to acquire backup CAS lock")

	// ErrCDCEnabled is returned when CDC is already enabled.
	ErrCDCEnabled = errors.New("CDC already enabled")

	// ErrInvalidVacuum is returned when the requested backup format is not
	// compatible with vacuum.
	ErrInvalidVacuum = errors.New("invalid vacuum")

	// ErrLoadInProgress is returned when a load is already in progress and the
	// requested operation cannot be performed.
	ErrLoadInProgress = errors.New("load in progress")

	// ErrNothingNewToSnapshot is returned when a snapshot is requested but there
	// are no new log entries to snapshot.
	ErrNothingNewToSnapshot = errors.New("nothing new to snapshot")

	// ErrNodeNotFound is returned when a node with a given ID is not found in the cluster.
	ErrNodeNotFound = errors.New("node not found in cluster")

	// ErrClusterNotFound is returned when a cluster should exist but does not.
	ErrClusterNotFound = errors.New("cluster not found")

	// ErrNoWALToSnapshot is returned when a snapshot is requested but there is no
	// WAL data to snapshot. This can happen when the Raft log contains only entries
	// that don't modify the database (e.g. cluster membership changes).
	ErrNoWALToSnapshot = errors.New("no WAL data available for snapshot")
)

const (

View on GitHub (pinned to 7586a4d1bd)

Solutions

1. Delay snapshot/backup operations until the load/restore completes
2. Serialize operational tasks: pause auto-snapshot schedules during planned restores
3. Retry the snapshot after the load finishes
4. Check load/snapshot status via node status endpoints before triggering either operation

Example fix

// before
// cron: run snapshot every minute, overlapping with /db/load
// after
if err := s.Snapshot(0); err != nil {
    if errors.Is(err, store.ErrLoadInProgress) {
        time.Sleep(30 * time.Second); return s.Snapshot(0) // retry after load completes
    }
}

// don't snapshot while a load/restore is running
if loadInProgress.Load() { return errors.New("load in progress; defer snapshot") }

func isLoadInProgress(err error) bool { return errors.Is(err, store.ErrLoadInProgress) }

if err := s.Snapshot(trailing); err != nil {
    if errors.Is(err, store.ErrLoadInProgress) {
        <-loadDone // wait for restore completion, then retry
        return s.Snapshot(trailing)
    }
    return err
}

When it happens

Trigger: Calling Store.Snapshot (or triggering an auto-snapshot) while a /db/load or /boot restore operation holds the load lock; the underlying raft.Snapshot() error mentioning "load in progress" is matched by string and converted to this error.

Common situations: Scheduled snapshots or auto-backup firing during a restore window; operator running /db/load while monitoring tooling requests a snapshot; retry logic re-triggering a snapshot mid-load.

Related errors

• acquiring read lock: %w
• failed to reload Raft config: %s
• ErrQueueClosed
• ErrCASConflict
• ErrCASConflictTimeout

AI-assisted analysis of rqlite/rqlite@7586a4d1bd (2026-09-03). Data as JSON: /api/errors/7fac66889ec5a676. Report an issue: GitHub.