lima-vm/lima · error

failed to check autostart registration: %w

Error message

failed to check autostart registration: %w

What it means

While launching the host agent, Lima checks autostart.IsRegistered to decide whether to spawn the host agent in the foreground or let the autostart manager do it. A backend error other than autostart.ErrNotSupported aborts the start with this message.

Source

Thrown at pkg/instance/start.go:254

	prepared, err := Prepare(ctx, inst, guestAgent)
	if err != nil {
		return err
	}

	if limactl == "" {
		limactl, err = os.Executable()
		if err != nil {
			return err
		}
	}
	haStdoutPath := filepath.Join(inst.Dir, filenames.HostAgentStdoutLog)
	haStderrPath := filepath.Join(inst.Dir, filenames.HostAgentStderrLog)

	begin := time.Now() // used for logrus propagation
	var haCmd *exec.Cmd
	if isRegisteredToAutoStart, err := autostart.IsRegistered(ctx, inst); err != nil && !errors.Is(err, autostart.ErrNotSupported) {
		return fmt.Errorf("failed to check autostart registration: %w", err)
	} else if !isRegisteredToAutoStart || launchHostAgentForeground {
		if err := os.RemoveAll(haStdoutPath); err != nil {
			return err
		}
		if err := os.RemoveAll(haStderrPath); err != nil {
			return err
		}
		haStdoutW, err := os.Create(haStdoutPath)
		if err != nil {
			return err
		}
		// no defer haStdoutW.Close()
		haStderrW, err := os.Create(haStderrPath)
		if err != nil {
			return err
		}
		// no defer haStderrW.Close()

View on GitHub (pinned to dd909d0973)

Solutions

1. Ensure a systemd user session exists (systemctl --user status works) and enable lingering: loginctl enable-linger $USER.
2. Run limactl from the same user session that owns the instance directory.
3. On truly autostart-less platforms, the ErrNotSupported path is tolerated - if you get this error instead, the backend exists but is failing; fix the session/dbus environment.
4. Alternative: limactl start --foreground (or the equivalent) to avoid the autostart decision path if available.

Example fix

// before
ssh host "limactl start myvm"  # no user session
// after
ssh host "loginctl enable-linger $USER && systemctl --user status && limactl start myvm"

// confirm a user service manager is reachable before start
if err := exec.Command("systemctl", "--user", "is-system-running").Run(); err != nil && runtime.GOOS == "linux" {
  log.Println("no systemd user session: start via autostart check may fail")
}

err := instance.Start(ctx, inst, "")
if err != nil && strings.Contains(err.Error(), "check autostart registration") {
  // retry in foreground path or after fixing the user session
  err = startForeground(ctx, inst)
}

When it happens

Trigger: `limactl start <inst>` on a host where the launchd (macOS) or systemd user (Linux) query fails: no systemd user session, dbus unavailable, or launchd plist lookup error.

Common situations: Starting instances from CI/containers/SSH sessions without a user service manager; running limactl as a different user than the one with the session; minimal Linux installs without systemd.

Related errors

• failed to check if the autostart entry for instance %#q is r
• failed to check if the autostart entry for instance %#q is r
• failed to %s the autostart entry for instance %#q: %w
• failed to unregister the autostart entry for instance %#q: %
• failed to check if the autostart entry for instance %#q is r

AI-assisted analysis of lima-vm/lima@dd909d0973 (2026-09-01). Data as JSON: /api/errors/9e07adb320216f13. Report an issue: GitHub.