Would a proxy appliance route DHCP broadcast traffic over VLAN borders?

I haven’t tested that ​@ChrisVe ...but that should be the case, thus why you need to add certain servers an an “Application Group” → i.e. DCs, DNS, DHCP, etc. Hopefully others have tested that out and provide you a definitive answer, but I do think that is the case.

From the Guide on Application Groups:

https://helpcenter.veeam.com/docs/vbr/userguide/application_group.html?ver=13

“When a SureBackup job is launched, Veeam Backup & Replication first starts in the virtual lab machines from the application group in the required order and performs necessary tests against them. This way, Veeam Backup & Replication creates the necessary environment for the verified machine. Only after all machines from the application group are started and tested, Veeam Backup & Replication starts the verified machine in the virtual lab.

For example, if you want to verify a Microsoft Exchange Server, you need to test its functionality in cooperation with the domain controller and DNS server. Subsequently, you must add to the application group a virtualized domain controller and DNS server. When Veeam Backup & Replication runs a SureBackup job, it will first start and verify the domain controller and DNS server in the virtual lab to make verification of the Microsoft Exchange Server possible.”

Hope that answers your question.

Thank your answer, but this doesn’t match my situation as I’m performing a DataLab test with VRO.

Ah...dang it. You did state that in your original post. Sorry Chris! Should be the same in VRO though. Let’s see if ​@AlecKing can give you a definitive answer.

​@Madi.Cristil ​@safiya ...it’s probably best to place this question in the VRO Hero’s Den. Thank you!

Hi! Yes, assuming that the configuration of the Virtual Lab appliance is correct and all VMs in the lab can reach the DC + DHCP server, then this should work. Remember to add the VRO step “Prepare VM as Domain Controller” to the DC specifically, this customises the restore commands so that DC will reboot into authoritative mode in the lab. 

Thanks ​@AlecKing for your statement. That makes me very relieved.

Hi,

If your servers are in different VLANs, they likely have different subnets attached to them already and they are different networks. The trick here is that DHCP by definition can’t traverse layer 3 networking because it has no IP or routes. So how is this built today? Do you have your DHCP server with one or more VLAN specific interfaces? If so you should recreate these within your isolated network, or if you need to reconfigure ranged, you’ll need to flatten your VLAN Topology as part of your VM Network Mapping configuration

​@MicoolPaul Oh my gosh! Thank you for opening my eyes. That’s what I have overseen the whole time. So, I have to use the built-in DHCP service of the virtual lab. Am I right, that the DNS I configure is updated automatically after some time?

Hey,

So you don’t “have” to use DHCP in your virtual lab, but you will need to make sure you’re spinning up your DHCP server for the duration of your testing (this is likely but doesn’t have to be your Domain Controller), which then means putting a “leg” in each virtual network. You can use DHCP but that will mean your servers will get different IP addresses and any static reservations made within your DHCP server will be ignored so you’ve got to scope your ranges. You can supply your own DNS Server configuration to the DHCP server so it can still route any DNS requests that are cross-subnet to the correct server(s) 🙂 hope this helps!

That’s what I understand from your previous post. As today the DHCP server has only one vNIC in production I have to add additional vNICs (one for each VLAN) to provide DHCP for that specific range. Right? But related to this have to make sure that the restored VMs keep their MAC addresses from production, if the VM should use their DHCP reservations.

Does the DHCP server have a virtual NIC in each scope? Or are you using DHCP relay on your switches for example as of today? Because the virtual NIC in each scope is supported, but DHCP relay is not sadly

No, the DHCP server has one virtual NIC and DHCP relay is used on the physical switches. For my VRO context this leads into two possible ways. Either scripting additional vNICs into the restored VM for each VLAN or using the built-in DHCP service of the virtual lab relying on that the DNS server is updated with the new IP addresses.

Note: A DHCP server does not require a separate NIC for each VLAN/subnet. With DHCP relay (IP helper) configured on the VLAN gateways (switch/router), a single NIC can serve multiple DHCP scopes. Is DHCP relay configured in your environment?

Yes, in the production environment DHCP relay is used.

But during a DataLab test DHCP relay is not available. There the DHCP server is in need to have “one leg” in each VLAN. 

Hi ​@ChrisVe, I haven't done this in practice similar to ​@coolsport00. The proxy applaince has a different role I guess, and does not forward DHCP broadcasts between VLANs which is the question you are asking here: https://helpcenter.veeam.com/docs/vbr/userguide/vlab_proxy_hv.html?ver=13  

But you could try using static IP mapping in the lab. There are other options but might be complex for you: https://helpcenter.veeam.com/docs/vbr/userguide/vlab_ip_hv.html?ver=13 ​​