Well, the rumours were true. vSphere5.1 has an all-new backup system that’s called vSphere Data Protection (vDP). This release sunset the older vSphere Data Recovery (vDR) appliance that was first introduces back in the vSphere4.0 days. Doesn’t that sound like a long time ago?
On the surface vDP and vDR are very similar and you might be forgiven in thinking that this is a name change/rebranding exercise. You’d be wrong. For a start vDP will be available in all the editions of vDR, and has new engine backing it – EMC’s Avarmar technology. It supports dedupe and change-block tracking not just for backups, but also restores as well – and is managed by the web-client. There will be three editions of the vDP corresponding with different capacities of usable storage for backup (0.5TB, 1TB, and 2TB), and of course you’re not limited to single appliance – many appliances can be added to the system. The appliance itself has been beefed up quite a bit – so its now 4 x vCPU system with an allocation of 4GB of RAM based on 64-bit version of SLES 11.
If your familiar with vDR you be pleased to hear that the look and feel remains very similar – so selecting VMs and their containers much the same except its all done via the web-client. One nice aspect of that is how that influence file-level restores. If you know vDR there used to be software client that could be installed inside the VM that would allow you to restore files directly into the VM. You could say it was kind of self-service restore for VMs. The FLR client would mount a backed up .VMDK to drive letter inside the guest operating system, and you would simply locate the file and copy across to the original location. That was all well and good but it meant an installation had to be completed first – albeit just 100MB or so. Now that’s all in the web-client and permissions system controls whether the application owner can only see the backups of the VM they are managing or whether they can see other backups that have occurred within vDR. Application-owners get to the vDP via the web-client:
http://:8580/flr
and by default will be presented a window of their previous backups. This identification process of mapping the VM to vDP backup jobs just require VMware Tools to be running.
Another nice feature I like is the “lock” option. vDP has the same “retention” policy that vDR had.
Basically, this controls the FIFO (first-in-first-out) process of how long you keep your backups for. That directly impacts on the quantity of storage required to keep that history. vDP has this lock option that can stop a backup from falling of the retention policy. So say there was major change to your VM scheduled on Tuesday, you could lock the backup made of the VM on Monday, to always give you a restore point that would take it back to that state before the change – rather than it being aged of the list of potential restore points overtime.
Two years ago when I was speaking at VMUGs around the US, I was making the point that “Project Cloud” needed to offer DR/Backup solutions inside the “Virtual Datacenter” for folks to have confidence in putting production workloads in that layer – and that’s precisely what these technologies are all about doing. So its not just another funky marketing term, we’re building out the technology to make that vision a reality.