Comment 3 for bug 497036

Just for the record, Gustavo started instance and well over 40 minutes later captured logs and rebooted. This *should* have been enough time for ec2-wait-for-meta-data-service to timeout.