kubernetes/kops · error

error creating user: %v Output: %s

Error message

error creating user: %v
Output: %s

What it means

The User task's RenderLocal creates the account by exec'ing `useradd` with built args and capturing combined output; a non-zero useradd exit becomes this error with useradd's own diagnostics appended, so the account creation is reported as a task failure.

Source

Thrown at upup/pkg/fi/nodeup/nodetasks/user.go:101

	if e.Shell != "" {
		args = append(args, "-s", e.Shell)
	}
	if e.Home != "" {
		args = append(args, "-d", e.Home)
	}
	args = append(args, e.Name)
	return args
}

func (_ *UserTask) RenderLocal(t *local.LocalTarget, a, e, changes *UserTask) error {
	if a == nil {
		args := buildUseraddArgs(e)
		klog.Infof("Creating user %q", e.Name)
		cmd := exec.Command("useradd", args...)
		klog.V(2).Infof("running command: useradd %s", strings.Join(args, " "))
		output, err := cmd.CombinedOutput()
		if err != nil {
			return fmt.Errorf("error creating user: %v\nOutput: %s", err, output)
		}
	} else {
		var args []string

		if changes.UID != 0 {
			args = append(args, "-u", strconv.Itoa(e.UID))
		}
		if changes.Shell != "" {
			args = append(args, "-s", e.Shell)
		}
		if changes.Home != "" {
			args = append(args, "-d", e.Home)
		}

		if len(args) != 0 {
			args = append(args, e.Name)
			klog.Infof("Reconfiguring user %q", e.Name)
			cmd := exec.Command("usermod", args...)

View on GitHub (pinned to 4c8573c808)

Solutions

1. Read the Output section — useradd messages like 'UID 1001 is not unique' or 'already exists' point to the exact conflict
2. If the user already exists with the right properties, reconcile state (remove and let nodeup recreate, or fix the task's Find/actual-state detection) and re-run
3. Choose a non-conflicting UID/Name in the cluster spec (InstanceGroup/userData) and re-run nodeup
4. Ensure the image ships the shadow-utils package providing /usr/sbin/useradd
5. Check /etc/login.defs constraints if useradd complains about ranges or home dirs

Example fix

// before: cluster spec
UserTask: Name "jenkins", UID 1000  # 1000 taken by image default user
// after
UserTask: Name "jenkins", UID 1100  # free UID on the image

// Pre-check UID availability and useradd presence before nodeup:
command -v useradd >/dev/null || echo 'useradd missing'
id -u jenkins >/dev/null 2>&1 && echo 'user exists'
getent passwd 1100 >/dev/null && echo 'UID taken' || echo 'UID free'

if err := task.RenderLocal(...); err != nil {
	if strings.Contains(err.Error(), "error creating user") {
		// Output embedded in error says why (exists / UID conflict)
		return fmt.Errorf("useradd failed; reconcile user state: %w", err)
	}
	return err
}

When it happens

Trigger: exec.Command("useradd", buildUseraddArgs(e)...) returns error: UID/GID already in use, username already exists, invalid -u/-g values, home directory conflicts, or useradd binary missing (e.g. on distros using adduser only).

Common situations: Cluster spec pins a UID that collides with an image-baked user, a previous partially-failed nodeup left the user created so a rerun hits 'already exists', or minimal images (some Debian slim/alpine-style) lacking useradd.

Related errors

• error executing command %q: %v Output: %s
• error doing systemd daemon-reload: %v Output: %s
• error doing systemd %s %s: %v Output: %s
• error doing 'systemctl %v': %v Output: %s
• error running kubectl: %v

AI-assisted analysis of kubernetes/kops@4c8573c808 (2026-09-05). Data as JSON: /api/errors/55f4b7c50aba0664. Report an issue: GitHub.