The Global Switch data center outage in the London Docklands earlier this month has been traced to a cable end box on a UPS system. The power interruption lasted less than a second, but it took at least one customer two days to fully recover.
The 222 millisecond outage was caused by a fault in a high-voltage cable end box of a circuit breaker on one of the facility’s diesel rotary UPS units, according to an incident report by Claranet reviewed by The Register.
For years, UPS system failure has been the most common cause of data center outages, according to Emerson Network Power-funded studies by the Ponemon Institute.
All Claranet customers lost access to services hosted in the GS2 data center, the report said. Claranet is a London-based managed hosting provider.
See also: Delta: Data Center Outage Cost Us $150M
Power was cut around 11am on September 10, but Claranet staff did not start powering its hardware back up for several hours, since they wanted to “ensure the facility was safe and that power supplies were reliable.” Only by around 5pm that day did they restore 85 percent of the services to customers.
“All components damaged by the HV fault have been replaced prior to normal power operations being restored,” Claranet’s report read. “The cable terminations of the remaining available DRUPS have been checked for integrity to ensure the same issue is not manifesting in other areas of the Global Switch 2 Power system.”
This was the second GS2 data center outage in three months, according to The Register. The first one, in June, also took out Claranet for several hours.