What causes "Restart script exited with non-zero exit status: %d" in hashicorp/packer?

Add the offending exit code to the accepted set by overriding restart_command to wrap the restart and exit 0 on that code (e.g. `...; if ($LASTEXITCODE -eq ) { exit 0 }`)

hashicorp/packer · error

Restart script exited with non-zero exit status: %d

Error message

Restart script exited with non-zero exit status: %d

What it means

The windows-restart provisioner runs the configured restart_command on the Windows guest and validates its exit status. Codes 1115 (ERROR_SHUTDOWN_IN_PROGRESS) and 1190 (ERROR_SHUTDOWN_IS_SCHEDULED) are accepted as successful reboot initiations; any other non-zero status means the restart command itself failed and Provision aborts with this message.

Source

Thrown at provisioner/windows-restart/provisioner.go:128

	p.cancelLock.Unlock()

	ui.Say("Restarting Machine")
	p.comm = comm
	p.ui = ui

	var cmd *packersdk.RemoteCmd
	command := p.config.RestartCommand
	err := retry.Config{StartTimeout: p.config.RestartTimeout}.Run(ctx, func(context.Context) error {
		cmd = &packersdk.RemoteCmd{Command: command}
		return cmd.RunWithUi(ctx, comm, ui)
	})

	if err != nil {
		return err
	}

	if cmd.ExitStatus() != 0 && cmd.ExitStatus() != 1115 && cmd.ExitStatus() != 1190 {
		return fmt.Errorf("Restart script exited with non-zero exit status: %d", cmd.ExitStatus())
	}

	return waitForRestart(ctx, p, comm)
}

var waitForRestart = func(ctx context.Context, p *Provisioner, comm packersdk.Communicator) error {
	ui := p.ui
	ui.Say("Waiting for machine to restart...")
	waitDone := make(chan bool, 1)
	timeout := time.After(p.config.RestartTimeout)
	var err error

	p.comm = comm
	var cmd *packersdk.RemoteCmd
	trycommand := TryCheckReboot
	abortcommand := AbortReboot

	// Stolen from Vagrant reboot checker

View on GitHub (pinned to eb36e3c3e4)

Solutions

Add the offending exit code to the accepted set by overriding restart_command to wrap the restart and exit 0 on that code (e.g. `...; if ($LASTEXITCODE -eq <code>) { exit 0 }`)
Verify the WinRM user has shutdown privileges on the guest
Test restart_command manually on the Windows instance to see the real error
Ensure no policy/AV software is blocking the shutdown command

Example fix

// before
provisioner "windows-restart" {
  restart_command = ["powershell", "-command", "Restart-Computer"]
}
// after
provisioner "windows-restart" {
  restart_command = ["powershell", "-command", "Restart-Computer -Force; if ($LASTEXITCODE -eq 3010) { exit 0 } else { exit $LASTEXITCODE }"]
}

Defensive patterns

Strategy: validation

Validate before calling

# Test the restart command manually on the guest before encoding it:
powershell -command "Restart-Computer -Force; exit $LASTEXITCODE"  # confirm exit codes 0/1115/1190

Prevention

Verify the WinRM user holds SeShutdownPrivilege and admin rights
Test custom restart_command exit codes on a disposable guest first
Wrap the restart command to normalize application-specific codes to 0
Check Group Policy/AV for shutdown blockers before provisioning

When it happens

Trigger: cmd.RunWithUi runs the restart_command (default: a PowerShell shutdown/restart) and it exits non-zero with a code other than 1115/1190 — e.g. access denied, shutdown.exe unavailable, or a custom restart_command that returns its own failure codes.

Common situations: Custom restart_command scripts returning application-specific error codes; insufficient Windows user privileges (SeShutdownPrivilege missing); antivirus or Group Policy blocking shutdown; typos in a custom restart_command.

Related errors

AI-assisted analysis of hashicorp/packer@eb36e3c3e4 (2026-09-05). Data as JSON: /api/errors/39a6754e24df9d9a. Report an issue: GitHub.