Possible to setup cross site backups using each site's repo?

[davidkillingsworth]

Hello,

We are using Veeam 9.5.

We have Site A
2 x ESXi hosts
1 x Veeam server is connected to a Synology NAS via iSCSI for it's local repository. Local backups are being stored in this repository.

Site A and B are connected via Site to Site VPN and each site has 100Mbit fibre broadband.

We have Site B
2 x ESXi hosts
1 x Veeam server is connected to a Synology NAS via iSCSI for it's local repository. Local backups are being stored in this repository.

We are currently copying the backups offsite via CloudBerry Labs to AWS.

This setup has been working well for the past several years without incident. The issue that the AWS S3 storage charges are getting very expensive. We also have tried a full VM guest image restore and it took 3 days to download the file as a disaster recovery test. This is basically wasted money at AWS if it takes a full 3 days to download a backup file to restore a single VM guest.

We have purchased extra storage for the Synology's at each site and would like to backup Site A's backups to Site B's repository, or Site B's Synology. We would also like to do the reverse. We would like to backup Site B's backups to Site A's repository, or Site A's Synology.

I have been reading forum posts from this forum and around the Internet for the past couple of hours before posting this, and it seems like the best way to do this is by using a Backup Copy Job to copy the backup files from the source to the target.

What I can't figure out is how we setup the source veeam server to the remote repository.

1) Is there a way to connect Site A's Veeam server to Server B's repository since it is already connected to Site B's Synology via iSCSI?
2) Do we need to map a drive from the Site A's Veeam server to Site B's Synology using SMB and then add that as a repository?
3) I am assuming that iSCSI is not a good protocol to use for connecting to the remote Synology because if Site A disappeared, then accessing the iSCSI LUN would not be possible, since the drive would have been formatted and mapped on the original Site A Veeam server. Do I have that right?
4) Is there any other method besides Backup Copy Job to do this, while keeping budget in the SMB/SME company price range?

The end result would be that we want to be able to restore Site A's VMs to new hardware in Site B in the event that Site A was totally lost due to flood or fire.... or be able to restore Site B's VMs to new hardware in Site A in the event that Site B was totally lost due to flood or fire.

[Andreas Neufert]

Hi David,

thanks for the request.
You can add the Repository Server on the other site to the B&R Server. Do the same on both sites.

Create a Repository on another folder, so that only one B&R Server can write to a specific folder.

Examples

Site 1 B&R perform backups to site 1 Repository on drive E in folder "RepositorySite1"
Site 2 B&R perform backup to site 2 Repository on drive E in folder "Repository Site2"

You add on site 1 Repository the folder "RepositorySite2BackupCopy" and on sitge 2 Repository the folder "RepositorySite1BackupCopy"

Then you add the Repository Servers to the other site and use the new created folder as target. So the BCJ will transport things accross the Server.

Repository roles running on servers can be shared as long as it runs on the same Backup & Replication build.

The only Veeam role in that context that do not support a shared scenario is the Veeam WAN accelerator source side role. But based on your feedback you do not use our WAN accelerator roles anyway.

Important as well is that the B&R Servers are not aware of each others. So you need potentially more CPU/RAM ressources as potentially now 2 B&R Servers write in parallel to the target storage.

[davidkillingsworth]

Then you add the Repository Servers to the other site and use the new created folder as target. So the BCJ will transport things accross the Server.

How do you do this part. I think that's the crux of my question.

I assume the following:
Site 1 Synology has a shared folder created specifically for Site 2's backups. This is called (Site1-Site2Daily).
Site 1 Veeam Server has that Synology share (Site1-Site2Daily) mapped in Windows on Drive: G.
Site 1 Veeam Server has Drive G: added as a repository, which is called (REPO-Site1-Site2Daily).

To allow Site 2's Veeam server to copy Site 2's daily backups via "Backup Copy Job," it needs to add the target location at Site 1.

Site 2 Veeam Server adds a new repository called (REPO-Site1-Site2Daily)
BACKUP INFRASTRUCTURE
ADD REPOSITORY
Call it (REPO-Site1-Site2Daily)
TYPE: MICROSOFT WINDOWS SERVER
REPOSITORY SERVER: ADD NEW
Type in Site 1's Veeam Server (add IP address, credentials, etc.)
POPULATE
Choose the G: drive of Site 1's Veeam server.

Do I have this right?

Do I actually need to define this step "Site 1 Veeam Server has Drive G: added as a repository, which is called (REPO-Site1-Site2Daily)" if Site 2's Veeam server is adding that G: drive as the repository? It seems like the G: drive on Site 1's Veeam server is added once to both Site 1 and Site 2 Veeam server. Is that the correct?

What happens if Site 2 disappears due to flood, fire, earthquake, etc.? Let's say I purchase a new physical server and install VMware vSphere on it in Site 1, but I want to restore all of Site 2's VM guests to it. I will go to Site 1's Veeam server and add the new VMware vSphere server, and then restore all of Site 2's VMs to new VMware vSphere server from Site 1's (REPO-Site1-Site2Daily). Is that correct?

The only Veeam role in that context that do not support a shared scenario is the Veeam WAN accelerator source side role. But based on your feedback you do not use our WAN accelerator roles anyway.

Let's say that I do want to upgrade our license so that we can take advantage of the WAN accelerator. What is the impact on my theorized setup above?

Please advise.

Thanks,
David

[foggy]

Site 1 Synology has a shared folder created specifically for Site 2's backups. This is called (Site1-Site2Daily).
Site 1 Veeam Server has that Synology share (Site1-Site2Daily) mapped in Windows on Drive: G.
Site 1 Veeam Server has Drive G: added as a repository, which is called (REPO-Site1-Site2Daily).

Since this is a folder created specifically to store Site 2's backups, you do not need to add it to Site 1's Veeam B&R instance. Just add it to Site 2's instance as a new repository and point corresponding backup copy jobs to it.

What happens if Site 2 disappears due to flood, fire, earthquake, etc.? Let's say I purchase a new physical server and install VMware vSphere on it in Site 1, but I want to restore all of Site 2's VM guests to it. I will go to Site 1's Veeam server and add the new VMware vSphere server, and then restore all of Site 2's VMs to new VMware vSphere server from Site 1's (REPO-Site1-Site2Daily). Is that correct?

Yes, you would need to import the Site 2's repository into the Site 1's Veeam B&R instance first (since it knows nothing about these backups) and then perform the restore.

Let's say that I do want to upgrade our license so that we can take advantage of the WAN accelerator. What is the impact on my theorized setup above?

You would need to have two different pairs of accelerators - one pair for each of the backup servers.