StartCommand exits with no logged error after 10-minute default timeout when a slave broker legitimately waits for the lock under startAsync=true

## Problem

A broker started via `./activemq start`, configured with `startAsync=true` and running as a slave, disappears after roughly 10 minutes with nothing logged.

## Root cause

`BrokerService.DEFAULT_START_TIMEOUT` is 10 minutes. `StartCommand` calls `broker.waitUntilStarted()` with that timeout, then throws if it returns false:

```java
if (!broker.waitUntilStarted()) {
throw new Exception(broker.getStartException());
}
```

Before `startAsync` existed, this was never an issue, because `start()` itself blocked until the broker became master, so `waitUntilStarted()` returned almost immediately either way.

With `startAsync=true`, `start()` returns right away, and a slave broker can legitimately wait on the lock far longer than 10 minutes — that is expected behavior, not a failure. When the timeout is hit, `waitUntilStarted()` returns `false` with no start exception set, and `StartCommand` treats that as a startup failure and exits the process. Because there's no recorded exception, nothing is logged, which makes the process exit look silent and unexplained.

## Environment

- `startAsync=true`
- Broker running as slave, waiting on the lock beyond the default 10-minute start timeout

## Suggested fix

`StartCommand` should keep waiting as long as `broker.getStartException()` is `null`, and only throw/exit when there is an actual recorded start exception.