Ads by TechWords

See your link here
Receive the latest technology news and information.
Computerworld Daily News (First Look and Wrap-Up)
Computerworld Blogs Newsletter
The Weekly Top 10
Cloud Computing
View all newsletters




Privacy Policy
 

About 165,000 Web sites knocked offline by NaviSite outage

The Web hosting company blamed a failed data center migration

November 6, 2007 12:00 PM ET

Computerworld - Approximately 165,000 Web sites have been offline since Saturday, thanks to a failed data center migration involving Andover, Mass.-based Web hosting company NaviSite Inc.

The problems started Saturday when NaviSite attempted to migrate and replace hundreds of servers operated by Baltimore-based Alabanza Corp., a Web hosting company acquired by NaviSite in August.

According to NaviSite spokesman Rathin Sinha, NaviSite decided to physically move 200 of the 850 servers operated by Alabanza to NaviSite's data center in Andover and then virtually migrate the data from the rest of the older servers to new boxes, also in Andover.

NaviSite let its customers know that their sites would be down for a while on Saturday, with the migration expected to be finished that day, Sinha said. But when NaviSite attempted to transfer the data from the 650 servers still in Baltimore it ran into a number of synchronization failures that kept multiplying.

As Saturday progressed, NaviSite realized it would probably miss its completion deadline; as a result, company officials decided to physically transfer another 200 servers from Baltimore to Andover to help reduce the scope of the virtual migration and speed up the data transfer.

But then NaviSite ran into more problems. As the hosts came up, their URLs did not, so although customers could access their Web sites from their IP address, they could not do so using their URLs, Sinha said.

"That was unanticipated," he said.

As NaviSite tried to solve that problem, the network became overloaded because of all the customers trying to get online, Sinha said. "What happened was first the URL could not match with the IP address and then IP did not match with the machine, so it took some time, and all this time we have a highly trafficked overloaded network," he said. "If there is one little problem, they multiply because there is a lot of dependencies."

Although Sinha said a "big chunk" of sites are back online, he could not say when everything might be back to normal. He also couldn't say how much this failed migration would cost -- NaviSite is a publicly-traded company.

To put it mildly, one of NaviSite's customers, Cynthia Brumfield, president of Emerging Media Dynamics Inc., an analyst firm in Washington, seems to be furious.

In an interview, Brumfield said she's going into her fourth day without access to her Web sites. And she said she doesn't believe the way NaviSite is spinning the story. While NaviSite said it has brought a large number of Web sites back online, she claims it hasn't.

"According to people who have talked to NaviSite's tech personnel, they were ill equipped for the relocation and ignorant of how to accomplish even basic tasks," she said in a blog post. "At this point, NaviSite's poorly planned data center consolidation has slipped from mere incompetence to outrageous indifference to its customers' needs and should be grounds for legal action, if not government sanctions of some kind."



Jump to comments

NaviSite

Additional Resources

EFD vs. HDD - What You Need to Know
WHITE PAPER
Enterprise flash drives provide a new Tier 0 storage layer capable of delivering high I/O performance at a very low latency. Proper use of EFDs in an Oracle environment can deliver increased performance compared to fibre channel drives. Read the recommendations for identification of the best DB components for EFDs.
Gartner Research Report: Magic Quadrant for Application Delivery Controllers, 2009
WHITE PAPER
The market for products to improve the delivery of application software over networks remains dynamic and innovative. Vendors focused on solving enterprises' most-pressing application problems have become the top players.
Eight Criteria for Server Load Balancing
WHITE PAPER
Server load balancers are a simple yet highly effective means to scale an application environment while ensuring its availability. Today's solutions should also address application performance and security. Read about the top eight criteria you should consider when choosing a server load balancer and how Citrix NetScaler meets those requirements.

What People Are Saying

White Papers & Webcasts

A Green Architectural Strategy That Puts IT in the Black
Levergage green computing across your data center. Read more now.  

Effectively Implementing Datacenter Automation
Effectively select and deploy the best datacenter automation solution today!

Master Data Management Projects in Practice - An Information Difference Research Study
Information Difference conducted a survey of both end-user organizations and systems integrators aimed at gaining deeper insight into MDM implementations and their success...  

XenApp Extends Virtualized Application Delivery
Download this webcast to learn how to accelerate delivery of virtualized applications and streamline management.

Open Source Master Data Management: The Time is Right
MDM is a natural extension to data integration and data quality. Open source MDM introduces a new, more accessible approach. It reduces implementation...  

Top HPC Use Cases in Life Sciences
Learn from the experts how best to apply cutting edge high-performance computing techniques a life sciences environment.


IT Jobs