Talk to the average solutions provider, and chances are they’ll be more than happy to sell you an expensive proprietary tool for VMware disaster recovery. Talk to VMware and they'll be more than happy to sell you Site Recovery Manager,
Defining your own VMware DR strategy
Most VMware admins know they need a disaster recovery strategy to get their VMs up and running in the shortest possible time if something goes wrong, but they’re not always clear on what tools they actually need. Don't let the solutions resellers define what you need, because you’ll probably end up paying way too much to protect a small VMware environment.
So what do you need for a VMware disaster recovery strategy in smaller environments? It depends on the number of data center sites you have, the amount of ESXi hosts you have, and the number of virtual machines (VMs).
Very small environments. If you have only one site, a single host and a few VMs, you may be able to get away with just having a decent backup strategy instead of full-blown DR. After all, if your site burns down, for instance, there is no place for the employees to do their work anyway.
Still, environments that are this small would benefit from hiring a recovery center that has hardware, office space and tape drives to help you resume business the day after a physical disaster. If you don’t have the ability to recover data in a different physical location, it makes no sense to purchase an expensive VMware DR solution in the first place. You could also employ a cloud to restore operations remotely after a disaster.
Multi-site infrastructure. Now let's assume that your company is a bit bigger and has two or more locations. Your basic VMware disaster recovery strategy in this case should be to have an equal amount of ESXi hosts at both sites, and ensure that each site can service all VMs after a disaster.
In addition, it’s a good idea for all the hosts to have access to the same VMs. For quick recovery, you can put them on a replicated storage area network (SAN) or use other replication methods such as VMware’s host-based replication. That way, if either site burns down, you can be sure that all data is still available at the SAN on the other site – and start the VMs from the replicated SAN on either site.
In a small environment, you may not require High Availability and vMotion as part of your VMware DR strategy. If you can do a manual failover inexpensively, don’t stress about getting these vSphere features.
Every VMware infrastructure should have a disaster recovery strategy in place when disaster strikes, no matter how small the environment or the disaster. For small companies, that may mean having off-site backups and a contract with a recovery site. For medium-sized companies, a replicated SAN may be enough to keep VMs and hosts running after a disaster.
This was first published in December 2011