There are just certain things in IT or the data center industry you don't want to hear - and fire and 'data loss' are two of them. There were a couple of lessons learned from fire and data loss stories recently and I post them here for just that reason - to learn from them.
The University of North Carolina Greensboro had one heck of a Friday last week. In a little under two hours they experienced two fires in the McNutt data center. After the first fire they were unable to determine the cause and staff investigating a short while later witnessed a server burst into flames, starting the alerts and emergency procedures all over again.
The root cause of the events was traced to a single server that overheated and likely burned up a wiring harness first (which produced the first event) and then continued to run until it blew up its power supply (which produced the second event). This was a new server that was not under load, and we believe this was caused by a manufacturer’s defect.
See their site for complete details of the event.
More of the Data Center Links post from John Rath