Where to place VBR/VRO?

[nhuotari-HBS]

Hello!

We are looking to possibly use VRO for our DR plan to restore vSphere VMs to Azure.

We have our VBR server on-prem on our vSAN cluster, which backs up our VMs to a SOBR with an onsite repo as the performance tier and then Azure Blob as the capacity tier.

My understanding is that we can use VRO to recover from the backups that are in Azure Blob much faster than we could on-prem to Azure.

Now, I am having troubles finding the answer to where to place the different components. If I'm not mistaken, the VBR and VRO servers should be in Azure with proxies doing the heavy lifting on-prem? My thought is that if we need to trigger our DR plan, it makes most sense to have the VMs that are going to facilitate that process, be running in our DR site (Azure).
If this is not correct, please let me know.

Now, assuming those VMs should be running in Azure, are there any other VMs that also need to be running in Azure (proxies, helpers, etc)?

Thank you in advance!

[skitch210]

With Azure Blob being a Capacity tier, you will need to put the performance tier into maintenance mode in order to pull directly from the Azure blob for the restores. You could setup VBR and VRO in Azure and have proxies on-prem, but you will need direct network access to on-prem from Azure, like Azure ExpressRoute or a VPN. You will want a very reliable network.

From a DR perspective, this would be the best location so the VBR and VRO are available in a DR event but you do run the risk if the network connection is offline at the time the backup is scheduled, the backup will not run.

[nhuotari-HBS]

Is that something VRO would be able to do for testing / activating a DR plan? (putting performance tier into maintenance mode).

Otherwise, would it be possible to break apart the SOBR and create a copy job to the Azure Blob and map to the existing backup files in the Blob?

[skitch210]

You could setup a custom Powershell script to put the SOBR into maintenance mode. You also could look to change your design to use a Blob as a Direct to Object copy job. The benefit to the copy job idea is that you can specify in VRO Recovery Location the Repository to use and tell it to be the Azure Blob.

[nhuotari-HBS]

Would it be possible to break apart the SOBR and map the new backup copy jobs to the existing backups in the Blob? We'd hate to have to seed everything again to a new Blob.

[Alec King]

I would agree with @skitch210 , you could use a PowerShell script to put the performance extents into maintenance mode. This script could be run in Orchestrator as a pre-plan step, so you can capture the output in VRO reports.

For the moment that's the only way to force capacity tier to be used - however we are already developing a method for VRO to force capacity tier (as a setting on Recovery Location), just for this kind of scenario. I can't promise which version of VRO it would ship in of course, but it is in progress

[nhuotari-HBS]

That is great news!

[skitch210]

I am testing out a script for this currently - hope to have it uploaded to Veeam Github early next week.

[nhuotari-HBS]

@skitch210, any updates on that script?

[skitch210]

Sorry for the long wait on follow up. This script is out in the Veeam Git-hub Orchestrator repository.
https://github.com/VeeamHub/veeam-orche ... entDisable
https://github.com/VeeamHub/veeam-orche ... tentEnable