weaviate/weaviate · error

file snapshot store: %w

Error message

file snapshot store: %w

What it means

raft.NewFileSnapshotStore creates the directory and retention policy for RAFT snapshots (nRetainedSnapShots kept); failure is wrapped as 'file snapshot store: %w'. Snapshots are required for log compaction and fast restarts, so startup cannot proceed without it.

Source

Thrown at cluster/store.go:565

		return fmt.Errorf("mkdir %s: %w", st.cfg.WorkDir, err)
	}

	// log store
	st.logStore, err = raftbolt.NewBoltStore(filepath.Join(st.cfg.WorkDir, raftDBName))
	if err != nil {
		return fmt.Errorf("bolt db: %w", err)
	}

	// log cache
	st.logCache, err = raft.NewLogCache(logCacheCapacity, st.logStore)
	if err != nil {
		return fmt.Errorf("log cache: %w", err)
	}

	// file snapshot store
	st.snapshotStore, err = raft.NewFileSnapshotStore(st.cfg.WorkDir, nRetainedSnapShots, st.log.Out)
	if err != nil {
		return fmt.Errorf("file snapshot store: %w", err)
	}

	// tcp transport
	advertiseAddress := net.JoinHostPort(st.cfg.Host, fmt.Sprintf("%d", st.cfg.RaftPort))
	tcpAddr, err := net.ResolveTCPAddr("tcp", advertiseAddress)
	if err != nil {
		return fmt.Errorf("net.resolve tcp address=%v: %w", advertiseAddress, err)
	}

	bindAddress := net.JoinHostPort(st.cfg.BindAddr, fmt.Sprintf("%d", st.cfg.RaftPort))
	st.raftTransport, err = st.raftResolver.NewTCPTransport(bindAddress, tcpAddr, tcpMaxPool, tcpTimeout, st.log)
	if err != nil {
		return fmt.Errorf("raft transport address=%v tcpAddress=%v maxPool=%v timeOut=%v: %w", bindAddress, tcpAddr, tcpMaxPool, tcpTimeout, err)
	}
	st.log.WithFields(logrus.Fields{
		"action":                 "raft_tcp_transport",
		"raft_bind_address":      bindAddress,
		"raft_advertise_address": advertiseAddress,

View on GitHub (pinned to 75aa4b6d11)

Solutions

1. Inspect/fix permissions on the raft work dir and its snapshots subdirectory
2. Free disk space and check inodes (df -i)
3. Remove corrupted leftover snapshot files if the node can resync from peers
4. Ensure nRetainedSnapShots (snapshot retention) is a sane positive value in source/config
5. Restore the whole RAFT dir from a consistent backup rather than partial copy

// pre-check snapshot dir
snapDir := filepath.Join(workDir, "snapshots")
if err := os.MkdirAll(snapDir, 0o755); err != nil { exit }
if unix.Access(snapDir, unix.W_OK) != nil { exit }

if err := raft.NewFileSnapshotStore(workDir, nRetainedSnapShots, logOut); err != nil {
    log.Fatalf("file snapshot store: %v — check dir perms, disk space, corrupt snapshots in %s", err, workDir)
}

When it happens

Trigger: Cannot create/read the snapshots subdirectory under WorkDir; permission denied; disk full; too many/invalid existing snapshot files; invalid retention count.

Common situations: Snapshot dir partially deleted or corrupted after crash; read-only volume; inode/space exhaustion; backups restored without the snapshots dir permissions.

Related errors

• snapshot shard %v: %w
• get tenant status for %s/%s: %w
• create snapshot for shard %s: %w
• stat objects bucket dir for shard %s: %w
• hardlink unloaded shard %s: %w

AI-assisted analysis of weaviate/weaviate@75aa4b6d11 (2026-09-04). Data as JSON: /api/errors/093177a536f9f149. Report an issue: GitHub.