Cut cable brings Wikipedia down
Redundancy plan didn't work, under investigation
By Anh Nguyen | Computerworld UK | Published: 11:00, 07 August 2012
Wikipedia, the free online encyclopaedia, experienced an outage on Monday after a fibre cable between its two data centres was cut.
The Wikimedia sites outage started at about 14.15 BST, and were restored at 15.18 BST. Mobile site services took a little longer to restore, and were back up at about 16.35 BST, the organisation said.
Wikimedia operates two data centres in Ashburn, Virginia and Tampa, Florida. They are connected by two, separate fibre links for redundancy. The Ashburn data centre serves most of the sites’ traffic, and talks to the Tampa data centre for back end services, such as databases.
Related Articles on Techworld
In a blog post, CT Woo, director of technical operations, wrote: “We do operate two 10-g separate fibres between the data centres. We are now working with our network provider to determine how and why we were impacted by that fibre cut when we are supposed to have redundancy in our network.
“We are still waiting for their full report.”
During the outage, the IT team rerouted traffic to the Tampa data centre, bypassing the Ashburn site.
Although connectivity has been restored to the network provider’s links (the first at 16.35 BST and the second at 19.30 BST), Woo said that the organisation has not reverted traffic back to Ashburn yet “until we are comfortable with their fix”.
According to the Wikimedia website, it runs on dedicated clusters of Linux servers, mainly Ubuntu, with a few OpenSolaris machines.