NZ university automates virtual machine backup

Auckland University will be upgrade its disaster recovery (DR) regime so that copies of 800-plus virtual machines will be automatically transferred between its main data centre and its DR site.

But the review reached a different conclusion. “Without IT, we cannot operate the business,” Askew says. “You could not actually do it. With no network and no authentication we cannot teach.”

The new DR site went live in 2007, but during the time taken to implement the new facility the University had started to make extensive use of virtualisation for servers and storage.

“We wanted to copy virtual machines and the data they rely on to the remote site,” Askew says. A method of doing so was devised, but required manual intervention. With virtual machines proliferating and disaster recovery processes becoming more complicated, the University decided to automate its processes.

“One of the great caveats of DR is that you plan for none of your people to be left standing so anyone can restore operations,” Askew says. “With a manual system that gets harder. That’s why we wanted an automated process to integrate storage and virtual machines. It was the next logical step.”

The University is now working on automated tools that use VMware Site Recovery Manager to copy virtual machines to its DR site whenever they are altered. “They’ll all be sitting there in standby mode,” Askew says. In the event of a disaster, the images of these virtual servers will automatically be installed on the server farm at the DR site, supplanting any of the test and development activity the site conducts when not performing its main function.

“Our recovery time objective is less than four hours,” Askew says. “We’ve tested it and done it in two.”