clock menu more-arrow no yes

Filed under:

Yesterday's Wikipedia outage explained: data center fiber connections severed

New, 9 comments

Wikimedia has posted an update to users saying that it was a severed fiber connection that led to yesterday's outage, and that the restoration should progress transparently.

Wikipedia eclipse
Wikipedia eclipse

Yesterday, Wikipedia was down for just over two hours when the fiber connection between the company's Ashburn and Tampa data centers was severed, according to Data Center Knowledge. The outage began at 9:15AM Eastern Time and ended at 11:35AM, and Wikimedia's backup link was restored at 2:30PM. It's this secondary link that has Wikimedia curious — the company is paying its network provider for redundant connections, but that failover evidently didn't kick in. While the company waits for its network provider to give its full report, the damage hasn't been completely repaired yet. Director of Technical Operations CT Woo said on the Wikimedia blog: "we [will] not revert traffic back to Ashburn yet until we are comfortable with their fix." However, Woo assures users that the switch back to the Ashburn data center should be a smooth one, and that they shouldn't notice a thing.