Fire and power outage take Colt Milan data centre offline

What happened

On Monday 6 July 2015 a fire closed Colt's Milan Lancetti Data Centre in Milan, Italy, taking the company's cloud services offline for several hours. Colt said it had experienced utility power supply outages at the site, that the vast majority of customer services had been fully restored and that the utility power supply was stable. Social media posts showed the facility completely dark, with smoke rising from the building and firefighters on site.

A report cited in the coverage said the fire may have been caused by overheating in the building's power infrastructure, but this was not confirmed. Temperatures in Milan had reached about 38 degrees C in the days before the incident. The Tier III site has 4MVA of power available. Reported outage durations differed, with one trade publication citing nine hours and one affected user citing about two hours.

Contributing factors

Colt reported utility power supply outages. A possible overheating fault in the building's power infrastructure was suggested by reports but not confirmed. Local temperatures were high at the time, but the sources do not link them to the fault.

Lessons learned

1. Operators should verify that building power infrastructure, including switchgear and distribution equipment, is inspected and thermally monitored, particularly during heatwaves.
2. Fire detection and suppression in electrical rooms should be tested to limit damage and prevent a single fault from taking a whole facility offline.
3. Communication plans should give customers accurate and timely information about outage causes and durations, avoiding gaps that rely on social media posts.

More lessons on weather and external hazards, detection and monitoring, outages, backup supply and restoration, from every incident on the register.

Source

datacenterdynamics.com