With our guaranteed 15 minute disaster recovery and several years under our belt with our product, we’ve gotten an understanding of the top reasons for performing a disaster recovery and the list may surprise you:
- Changing storage platforms
- Failed ESXi data store
- Failed server
- Human error
- Moving data around the world
What? A true disaster isn’t on the list? There’s no hurricane, earthquake, etc. on there. That’s right folks, DR is rarely performed for a true natural disaster. Let’s walk through the list and talk about the reasons.
Changing storage platforms
Once your data is free from the encumbrances of the hardware you are freed to change your mind on how you deploy your storage. Customers often start with us using a virtual machine and then later decide to move to hardware to improve performance. In the old world, this would mean a massive data migration from one platform to another, but with Nasuni this is a 15 minute process. Customers move from VM to hardware, from ESX to Hyper-V and terabytes of data move with them, in just a few minutes. This is by far the #1 use case of DR with Nasuni.
Failed ESXi data store
While its frightening to say, this is the second leading cause for DR with Nasuni. The underlying storage behind your virtual infrastructure should have enough performance to meet your needs and be robust and protected. It’s rare that a week goes by with our growing customer base where we don’t see yet another problem with a data store going south. Most of the time customers have chosen to use less expensive iSCSI devices behind their servers and those systems are just not that well built. To many it’s common sense, but storage is important because it has state and the data is hard and expensive to recreate. While you should look for the best cost when selecting your storage provider, you also need to look at the cost to your business of losing your data or spending a day or more rebuilding your virtual infrastructure. Don’t put your virtual infrastructure on inferior storage platforms. We’re saying this not as one of the storage providers, but as one of the virtual machines running in your virtual infrastructure. There have been several cases where the poor storage went south and the customer quickly recovered their filer through the DR process but still had many hours ahead of them recovering the other virtual machines they had from backups.
Similar to the failing data store, customers sometimes pick older or lower quality hardware to run their virtual infrastructure. We’ve seen all sorts of failures in the servers themselves that have managed to take down many virtual machines with them. For both this and the previous cause, the best bet is to invest in quality hardware — for both your compute and storage needs. If needed, move your most important things onto dedicated hardware, the Nasuni Filer is a good choice for that. We have yet to see a DR needed because of failed hardware provided by Nasuni.
Virtualization gives IT tremendous power. You can move machines around within your internal cloud, re-balance workloads, do dynamic process migration and all other sorts of magic. But with power comes responsibility. We’ve seen cases where the overworked, underpaid IT folks were moving too fast and right-clicked on a data store and clicked “delete” and then even confirmed the action only to later realize they had deleted a major data store underlying several virtual machines on their system. At that point the data is gone. This is like losing the the disk from a failed data store but due to human error. Its a disaster and its embarrassing. With a fast DR IT can get the Nasuni Filer back quickly, but if there were other virtual machines on that data store they can be in for a long haul.
Moving data around the world
Number 5 on our list of the top reasons for DR goes back to the time before we had multi-site capabilities. We had customers that wanted to set up a remote office and they wanted to send terabytes of data to that remote office but they had no good way of getting all the data there and accessible. What they did was set up a virtual filer, load it with data then they shut it down and performed the DR at the remote office. In less than 15 minutes all of their data was instantly accessible even though the filer was now thousands of files from where it was created and updated only minutes ago. We saw several customers use this as a deployment model for remote offices until we provided the ability to access the same data globally. Now DR’s are not needed, but they still rank high on the all-time list of reasons for DRs.
Those are the top reasons DR’s happen with Nasuni. We hope to never see the day when a DR is needed for a true disaster from natural or human-caused events, but if we do Nasuni will be there to help get your business back up and running.