A few years ago, Information Systems Manager Howard Chason explored worst case scenarios -- such as a tornado destroying the city of Batavia, Ill.'s data center -- and didn't like what he learned. It would take two to four weeks to fully recover and get every application back up and running.
Those two to four weeks includes purchasing new servers and waiting for them to be delivered. It would also include reinstalling software and recovering data from backup tapes. He needed a much faster recovery time.
"A lot of the services we provide like GIS need to be back up immediately," he says.
So the city decided to invest about $100,000 in two new servers, VMware virtualization software and two 8TB SANs, so it could replicate its servers and data from the city's main data center to a second data center. Batavia made the purchases over two years in 2007 and 2008.
Today, the two servers in both locations split the workload and operate about 12 to 15 virtual servers. Snapshots of the data in the main SAN are replicated to the secondary SAN five times a day. If downtime occurs, Chason estimates that he and his IT staff could get the servers back up and running within 15 to 30 minutes.
For further data protection, the city still performs tape backup using LTO-4 technology and Symantec's Backup Exec software.
"Disaster recovery is an insurance plan," he says. "You hope you don't have to use it, but if you do, you are ready."