B&R Scaling up and out with Hyper-V Cluster

[qcnetworks]

Have been using Veeam B&R for a few years now and it's working pretty well. Started with about 10VM's to back up, using a Dell 740XD2. Now we are over 100 VM's and backups are slow. I've added a few Veeam VM's as Proxies, setup a few iscsi connections to a couple of QNAP/Synology devices...but now it's time to grow up. So, question, what is the best way to set up a B&R solution for 150 VM's? About 40 of these VM's are 2Tb in size and the rest are general 400-gb vm's. Should I go with 2 or 3 large servers with internal storage and ReFS, or 6 small servers with 4 drives and 50Tb each, or Veeam servers as VM's with iscsi connections to NAS devices? Or....Need to know what best practice is. Is the NAS ok to use (formatted as ReFS on windows vm's) or is it better with internal drives on physical servers? Which are more stable: physical (with internal storage, or even iscsi connections) or virtual Veeam servers (with iscsi connections)? What are you using for physical veeam servers (brand/model/specs would be good info). I need this to be scaleable to double in size as well, so 300 VMs in 2-3 years. I have one large Veeam server running Linux hardened repo with internal storage (100Tb)...seems to be working well. I'm nervous about putting all my eggs in 1 basket tho. Have been bit in the past. And...do you backup your backups? Do you have any kind of backup solution for your repos? If i have 8mo worth of chains and the server and repo falls over, I dont want to lose these backups. But now we are talking about the least expensive 500Tb solution I can find (to backup the backups). Thx in advance for you expert advise!

[Mildur]

Hi Scot,

What bottleneck is indicated in your backup job session details?

I'll try to answer some of your questions, but please note that we are limited in this forum regarding how far we can go in recommending a backup design. I suggest reaching out to a VASP Partner from Veeam to discuss and review your new backup design specific to your environment. They can also provide more information about specific hardware vendors and do the correct sizing for you: https://www.veeam.com/partners/find-a-p ... ?type=VASP

For backup targets, I recommend using Hardened Repositories (with built-in disks) over ReFS. There are currently some known issues with ReFS. If you prefer not to use two Hardened Repository targets, you can also consider options with object storage appliances.

Personally, I would exclude NAS devices as backup targets from my backup design. NAS devices with iSCSI ReFS volumes can work, but be sure to use enterprise-grade NAS devices, not those intended for home use. We don't recommend NAS devices for Hardened Repositories, as they introduce additional attack vectors for malicious actors. However, I am aware of a few customers who are successfully using enterprise NAS systems with FC connections to their Windows/Linux hosts.

And...do you backup your backups? Do you have any kind of backup solution for your repos? If i have 8mo worth of chains and the server and repo falls over, I dont want to lose these backups. But now we are talking about the least expensive 500Tb solution I can find (to backup the backups).

You can use Backup Copy Jobs to send data to a second repository. If you expect to have 500TB of backup data, it’s best to use two repositories: one for your backup jobs and another for your copy jobs. Don't forget to add enough space for unexpected events such as Active Fulls.
We don't recommend creating a "backup" of the entire repository. Backup Copy Jobs are sufficient to maintain a second, independent backup chain on another target, and potentially even in another location.

You may also want to check out our Best Practice Guide to answer some of your questions.

And please also take a look at our calculator for the storage calculation: https://www.veeam.com/calculators/simple/vbr/machines

Best regards,
Fabian

[qcnetworks]

Thank you Fabian!