UPDATE (July 7): An Equinix spokesperson has confirmed that the outage did not involve the Equinix facility that provides services to Wikimedia in Florida. A Wikimedia spokesperson was not immediately available for comment.

Original article from July 6th:

A Wikimedia Foundation data center in Tampa, Fla., experienced a power failure early Monday morning, bringing down every online property of the famous non-profit provider of the user-contributed and user-edited online encyclopedia Wikipedia.

The foundation uses two data center providers in Florida: Switch and Data (now owned by Equinix) and Hostway. It was not clear from the announcement of the downtime on the organization’s tech blog which provider’s facility went down shortly after midnight UTC on Monday. Power in the data center was returned back to normal about one hour later and engineers began the process of recovering systems and services, expecting all projects to be back online around 4 a.m. 

This is a second reported data center outage for Wikimedia. Its Amsterdam data center overheated in late March, forcing the servers inside to shut down automatically. A broken failover mechanism prevented a successful transfer to a Florida data center, resulting in an outage that lasted longer than one hour. 

Wikimedia’s IT infrastructure in Europe is hosted by EvoSwitch and SARA. 

The organization admitted the weakness in downtime protection of its infrastructure in its annual 2010-2011 plan and said it would add another data center to enable safe failover, in addition to improving monitoring, capacity planning and operations response. Wikimedia plans to spend $3.27m on the new site.

In February, Google gave Wikimedia a $2m grant, which the organization said it would use to expand its data center capacity, the first major task for its new CTO Danese Cooper, who took on the role in the beginning of the year.