feat: re-add daemonize to start.sh and loop waiting for ./readyz

[fkondej]

No description provided.

[claude]

Claude finished @fkondej's task in 1m 36s —— View job

---

Review of PR #391

• Gather context and read changed files
• Analyze code changes for issues
• Post review feedback

Summary

This PR changes start.sh from using exec (foreground QEMU replacing the shell) to -daemonize (QEMU forks into background), and adds a readiness polling loop that waits for the HAProxy /readyz endpoint to return 200.

The overall approach is sound. Two minor issues flagged inline:

1. Sleep-before-check pattern (lines 100-116) — the loop sleeps 5s before the first health check, adding unnecessary delay if the VM boots quickly. Suggest checking first, then sleeping.
2. Missing cleanup hint on timeout (lines 118-120) — when readyz times out, the QEMU VM is still running but the error message doesn't mention stop.sh. Users may be left with an orphaned VM.

No bugs, security concerns, or correctness issues found.

[claude]

The loop sleeps 5s before the first health check. If the VM is already ready (or becomes ready quickly), this adds an unnecessary delay. Consider checking first, then sleeping:

Suggested change

	while (( SECONDS < DEADLINE )); do
	sleep 5
	# Fail fast if QEMU died (bad image, kernel panic, etc.)
	if ! kill -0 "$(cat "${PIDFILE}")" 2>/dev/null; then
	echo "Error: QEMU process died during boot"
	echo " Check console log: tail -f ${CONSOLE_LOG}"
	exit 1
	fi
	HTTP_CODE=$(curl -s -k -o /dev/null -w "%{http_code}" "${READYZ_URL}" 2>/dev/null || echo "000")
	if [[ "${HTTP_CODE}" == "200" ]]; then
	echo "VM ready after $(( READYZ_TIMEOUT - (DEADLINE - SECONDS) ))s (PID: $(cat "${PIDFILE}"))"
	exit 0
	fi
	echo " waiting... $(( READYZ_TIMEOUT - (DEADLINE - SECONDS) ))s (HTTP ${HTTP_CODE})"
	done
	while (( SECONDS < DEADLINE )); do
	# Fail fast if QEMU died (bad image, kernel panic, etc.)
	if ! kill -0 "$(cat "${PIDFILE}")" 2>/dev/null; then
	echo "Error: QEMU process died during boot"
	echo " Check console log: tail -f ${CONSOLE_LOG}"
	exit 1
	fi
	HTTP_CODE=$(curl -s -k -o /dev/null -w "%{http_code}" "${READYZ_URL}" 2>/dev/null || echo "000")
	if [[ "${HTTP_CODE}" == "200" ]]; then
	echo "VM ready after $(( READYZ_TIMEOUT - (DEADLINE - SECONDS) ))s (PID: $(cat "${PIDFILE}"))"
	exit 0
	fi
	echo " waiting... $(( READYZ_TIMEOUT - (DEADLINE - SECONDS) ))s (HTTP ${HTTP_CODE})"
	sleep 5

[claude]

On timeout, the QEMU VM is still running in the background but the user isn't told. Consider adding a hint to run stop.sh:

Suggested change

	echo "Error: VM did not become ready within ${READYZ_TIMEOUT}s"
	echo " Check console log: tail -f ${CONSOLE_LOG}"
	exit 1
	echo "Error: VM did not become ready within ${READYZ_TIMEOUT}s"
	echo " Check console log: tail -f ${CONSOLE_LOG}"
	echo " Stop the VM: ./scripts/stop.sh"
	exit 1