Toigo, CEO and managing partner of Toigo Partners International, said new technologies and budget cuts shouldn't lull IT administrators into thinking they don’t need to make DR a top priority.
“The biggest noisemakers out there right now are the server virtualization and hypervisor vendors,” Toigo said during his talk on IT disaster recovery planning Tuesday. “They are trying to make it sound like you don’t need disaster recovery anymore.”Toigo also dismissed arguments that tape is an outdated technology as an integral part of a IT disaster recovery plan. “For those who say tape is dead, they’re full of it,” Toigo said. He said disk can get corrupted or fail, requiring tape to save the day.
Toigo gave several recommendations for efficiently protecting data now and getting your DR system on track. His first piece of advice was to do a successful backup now. Toigo said you’d be surprised how many organizations don't have successful and recent backups, mainly because of budget cuts and increasing workloads.
“Give yourself at least that baseline data that you can work from,” he said.
He also suggested that administrators should do more to document downtime events as teachable moments. “You can learn to operate your storage infrastructure more efficiently just by looking at what caused [an outage] and what can be done in the future to avoid it,” he added.
Recording what happened, why, and the downtime and economic consequences can also help you develop a fact-based argument for DR to take to management. Collecting statistics on your own company’s downtime will have more impact than using generalized industry statistics.
He also told the audience to get serious about using data archiving to pull little-used data off production equipment and out of routine data protection processes such as replication and backup. Toigo also urged attendees to make sure they understand all of the data that is required to recover an application.
He used Microsoft Exchange as an example, pointing out that not only would you need mailbox data, but also device drivers, mailbox configuration, and Active Directory settings. “The application isn’t just the sum of its data,” Toigo explained. “It’s all this other stuff. And you have to catch all of it if you are going to recover that environment.”
Toigo also exhorted his listeners to break their mirrors to confirm they are working because the DR guy might be the last to know when volumes get moved around and applications get removed and added. “Stop the mirror and do a consistency check, and see if what you think you’re mirroring is actually what you are mirroring,” he said. The wrong time to find out is when you have the disaster.”
Other suggestions included:
- Use storage virtualization to eliminate vendor lock-in and get local and regional interruption protection at the storage hypervisor level;
- Perform delta checks at primary and replication sites because WAN latency can jeopardize transaction and database data protection schemes;
- Stage 30 days of backups to disk to quickly recover from user-caused disasters such as accidental file deletion; and
- Turn your smaller DR site into a test lab for technologies you can’t get funding for in production environments. Toigo also mentioned virtual desktops and IP telephony as examples of those technologies.