Rapid Recovery in large Enterprise environments

Hi All,

I am coming from a SMB where we used RR very successfully to a huge environment where backups are frankly kind of a mess.

Does RR scale up well to large environment problems? For example instead of stand alone SQL DB's with maybe a few hundred GB's, Its SQL AG's of tens of Terabytes in size.

Instead of hyperV hosts we have VSphere clusters with hundreds of VM's each.

Does RR's features scale up? Can you protect environments of this size and scale reliably and use things like standby VM's and replication?

Parents
  • A few things to consider here. Generally when you start adding more and more VMs considering moving from agent based to agent-less might be a good move. Also once you do that you'd need to make sure that things like your virtual center and/or hosts have the adequate resources to cater to hypervisor integration. Rapid Recovery doesn't do a bad job scaling, as long as you scale along with it. You start adding more nodes, and more protection, larger repos, make sure the RR server itself gets more resources too (RAM and dedupe cache, CPUs). I have a couple of instances with RR protecting over 100 nodes per instance, and a handful dozens - if we give the cores enough RAM she does the job. 

    It does become a resource question, and currently without a virtual appliance in the RR line at this point, those resources would have to come from the RR core itself.

Reply
  • A few things to consider here. Generally when you start adding more and more VMs considering moving from agent based to agent-less might be a good move. Also once you do that you'd need to make sure that things like your virtual center and/or hosts have the adequate resources to cater to hypervisor integration. Rapid Recovery doesn't do a bad job scaling, as long as you scale along with it. You start adding more nodes, and more protection, larger repos, make sure the RR server itself gets more resources too (RAM and dedupe cache, CPUs). I have a couple of instances with RR protecting over 100 nodes per instance, and a handful dozens - if we give the cores enough RAM she does the job. 

    It does become a resource question, and currently without a virtual appliance in the RR line at this point, those resources would have to come from the RR core itself.

Children
No Data