Re-Use IP address for Backup Infrastructure server

[Taroking1]

Thought I would post about a recent problem we came across. This is not a Veeam issue per se, but could trip you up if you are attempting to re-use the IP address on a backup infrastructure server.

Scenario
We have Veeam servers in each of our remote locations that perform the function of Repository, Wan Accelerator.
We are replacing all of our current hardware out in each site and want to build a new Veeam server on the new hardware, migrate the repository and Re-Use the IP address of the existing server. We want both the new repository server and the old one online at the same time to allow us to use the Evacuate Repository feature.
We have a very very locked down environment with multiple firewalls. Just deciding to assign a new IP address to the new Veeam remote server will require a LOT of work and documentation and was deemed as our last option.
We have been allowed to use a temporary IP address to allow the old Veeam repository server and the new Veeam repository server to exist at the same time in the same firewall group in the remote site. This temporary address will also allow traffic from the VBR server and will only exist for 24 hours or so while the repository migration happens.

Our approach
1. OldRepoServer - used to have permanent IP address x.x.x.116
2. Change OldRepoServer ip address to temporary IP address of x.x.x.117
3. Confirm OldRepoServer is accessible in VBR console, Run a rescan. Everything working as expected.
4. Commission new Repo server with the original IP address - NewRepoServer with IP address x.x.x.116
5. Add NewRepoServer to VBR. VBR reports that all components are already installed. This is not correct. This is a brand new server, so at this point we know there is a problem.
6. If you now try and add NewRepoServer as a Backup repository server it shows all of the files and details from OldRepoServer. So problem confirmed

Problem
Veeam believes these two servers are the same item.
If you open Powershell from the VBR console and run a Get-VBRServer -name <servername> the ID, Uid and name are unique, except for the PhysHostId which shows as a duplicate. Now in this instance, that cannot possibly be right, this is a brand new server build via an automated build task sequence, so no imaging or anything.
So now the problem is identified, a duplicate PhysHostID. But how does it generate this PhyHostID?

Solution
AS far as we can see, Veeam generates the PhysHostId from the IP address when you add it as a part of your Backup infrastructure, so now the solution is actually very simple.
For us the first step then becomes, give NewRepoServer the temporary IP address and add it to Backup Infrastructure in VBR. The new PhysHostID will be generated at this time and VBR server will recognise this as a new server and install the correct components.
After this, we then swapped the IP addresses - waited for the machines to register their IP address change in DNS and it replicated to the VBR server site (very important) and then we are able to proceed with our repository migration

[HannesK]

Hello,
before going into details... can you maybe explain a little bit more step 2 and 3?

Is that about the IP address change on the OldRepoServer operating system? I'm assuming that you registered OldRepoServer with DNS in VBR?

If an IP address changes in a Windows environment with default settings, then flushing DNS cache is needed, yes. What I always do is:

1) change IP
2) ipconfig /flushdns on 3rd machine
3) check whether DNS resolution works
4) if not, re-register DNS (ipconfig /registerdns or reboot)
5) only continue if DNS resolution works

Best regards,
Hannes

[Taroking1]

HI HannesK

The old server is registered in DNS. The VBR server is able to resolve both servers on their new address. This is not a DNS issue.

The issue is as explained. Veeam generates the PhysHostId from the IP address when you add it as a part of your Backup infrastructure.

You can recreate by doing this
Give Server A and IP address of 10.1.1.1
Add to Veeam VBR and give is a role (repo, wan accelerator, anything)
Now change the IP address to 10.1.1.2 and force a DNS re-registration. Confirm from the VBR server that you can ping the name of ServerA and it has picked up the new IP address. Also rescan the server in VBR for good measure to 100% confirm the IP address change has been picked up.
Now build Server B and give it IP address 10.1.1.1
Add to Veeam VBR. Try to add any Veeam roles, notice that it says it has all of the roles you have assigned to Server A?

Fix the issue by removing Server B from Veeam. Change it's IP address to 10.1.1.3.
Add to Veeam VBR.
Now change the IP address of Server B to 10.1.1.1
Wait for the DNS re-registration to happen and VBR to recognise the server on it's new IP address.
Add roles to Server B as you wish.

[HannesK]

hmm, I just tried it again and it worked for me. I just did rescan of the server and then repository. I could add a new proxy on the new server. Worked fine. I never removed anything.

Maybe PowerShell is the easier way to go anyway. I forgot about that way before. The KB article should be updated as it works for all registered servers. Not only VCenter.