I left extremely strict instructions not to reboot a server which had been very susceptible to kernel panics on restart. Does anyone want to guess what the researcher did next?