Latest news of the domain name industry

Recent Posts

New gTLD registry hopeful CentralNic taken out by “total power failure”

Kevin Murphy, March 15, 2013, 18:34:15 (UTC), Domain Registries

Emerging new gTLD back-end player CentralNic today suffered a two-hour blackout of its registry systems, due to a “total power failure” at its data center.
Its registry, which handles subdomain services such as uk.com and gb.com and the ccTLD .la, was offline from 0930 to 1130 UTC this morning, the company said.
Even though the company has all the necessary backup precautions you’d expect from a total-uptime domain name registry, for some reason they failed to kick in, it seems.
CentralNic said:

The data centre is equipped with fully resilient power supply including N+1 redundant [Uninterruptible Power Supply] arrays and backup diesel generators, and the exact cause of the outage, and why the UPS and diesel generator system did not take over to maintain power, is not yet known.

The company is the named back-end provider for 60 gTLD applications, including a handful of dot-brands.
Coming so soon before ICANN starts the pre-delegation testing of registry providers, the outage is embarrassing to say the least.

Tagged: , ,

Comments (5)

  1. Gavin Brown says:

    We’ve just posted an update based on a report received from our hosting partners. I’ll copy it here for your readers.
    At 0945hrs UTC, the national power grid that supplies power to the Level 3 data centre at Goswell Road in London experienced an outage. This facility is Level 3′s flagship European data centre and is one of the largest in London. When the mains power failed, the facility automatically switched to emergency power generators, but an isolated UPS fault (which was not detected during the most recent weekly test of the UPS system) subsequently resulted in a loss of power to part of the facility. Unfortunately, this included the collocation suite which houses CentralNic’s primary operations centre.
    At 1100hrs UTC, Level 3 restored power, and we began service restoration. By 1115hrs UTC, core network systems had been restored, and most registry systems came back online shortly thereafter.
    Had registry services not been restored at this time, we would have initiated a failover to our Disaster Recovery site in the Isle of Man.
    We have been assured by Level 3 that this was an isolated incident and there is absolutely no way it could reoccur. For our own part, we will carefully analyze today’s event to ensure that all possible measures are put in place to avoid or mitigate any future re-occurrence.

    • Gavin Brown says:

      It should also be noted that no disruption to registrants resulted from this outage. Domain registration requests and updates are generally queued by registrars for submission to the registry once it is restored. And of course, at no time was the resolution of domains affected – our DNS system is spread out over almost fifty cities on four continents.

  2. MarkA says:

    Uh, Gavin?.. Isn’t that why we have hot back-ups and redundant fail overs? “And ‘you’ wanna be my latex salesman?!?”

  3. Gary says:

    How many outages has Verisign had in all their time?
    Enuf said.

Leave a Reply to Gary