kubernetes/kops · error

error doing systemd %s %s: %v Output: %s

Error message

error doing systemd %s %s: %v
Output: %s

What it means

RenderLocal applies the chosen systemd action (start/stop/restart, run with --no-block) via exec and includes the command's combined output in the error when systemctl exits non-zero. This means systemd itself failed to schedule/perform the action on the service.

Source

Thrown at upup/pkg/fi/nodeup/nodetasks/service.go:401

						klog.V(2).Infof("will restart service %q because dependency changed after service start", serviceName)
						action = "restart"
					} else {
						klog.V(2).Infof("will not restart service %q - started after dependencies", serviceName)
					}
				}
			}
		}
	}

	if action != "" && fi.ValueOf(e.ManageState) {
		args := []string{"systemctl", action, serviceName}
		// We use --no-block to avoid hanging if the service has issues stopping/starting
		args = append(args, "--no-block")
		cmd := exec.Command(args[0], args[1:]...)
		klog.Infof("Restarting service %q (running %q)", serviceName, strings.Join(cmd.Args, " "))
		output, err := cmd.CombinedOutput()
		if err != nil {
			return fmt.Errorf("error doing systemd %s %s: %v\nOutput: %s", action, serviceName, err, output)
		}
	}

	if changes.Enabled != nil && fi.ValueOf(e.ManageState) {
		var args []string
		if fi.ValueOf(e.Enabled) {
			klog.Infof("Enabling service %q", serviceName)
			args = []string{"enable", serviceName}
		} else {
			klog.Infof("Disabling service %q", serviceName)
			args = []string{"disable", serviceName}
		}
		cmd := exec.Command("systemctl", args...)

		output, err := cmd.CombinedOutput()
		if err != nil {
			return fmt.Errorf("error doing 'systemctl %v': %v\nOutput: %s", args, err, output)
		}

View on GitHub (pinned to 4c8573c808)

Solutions

1. Read the Output section of the error — it contains systemctl's own message; fix the unit or binary it names
2. Run `systemd-analyze verify /etc/systemd/system/<svc>` to validate the unit
3. Check the unit isn't masked: systemctl status <svc>; unmask if needed
4. Run the restart manually: systemctl restart <svc> to reproduce and see full logs (journalctl -u <svc>)

Example fix

// before: unit with bad path
ExecStart=/usr/local/bin/protokube   # binary not present
// after: ensure binary installed or fix path
ExecStart=/opt/kops/bin/protokube

// Validate before applying:
systemd-analyze verify /etc/systemd/system/<unit>
systemctl cat <unit> >/dev/null && echo 'unit loadable'

if err := task.RenderLocal(...); err != nil {
	if strings.Contains(err.Error(), "error doing systemd") {
		// error embeds action, service, and systemctl Output
		return fmt.Errorf("systemctl action failed; fix unit/binary per Output: %w", err)
	}
	return err
}

When it happens

Trigger: exec of `systemctl restart --no-block <svc>` (or start/stop) returns non-zero: unit file invalid after write, service binary missing, unit masked, or systemd rejecting the request (dbus down, not booted with systemd).

Common situations: A hand-edited or generated unit with bad ExecStart path, a masked unit, running nodeup on a system without systemd, or dependency ordering failures reported by systemd.

Related errors

• error doing systemd daemon-reload: %v Output: %s
• error doing 'systemctl %v': %v Output: %s
• error building kubelet flags: %v
• failed to create bpf mount unit: %w
• failed to create cgroupv2 mount unit: %w

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/e2d92b8fce1e336f. Report an issue: GitHub.