Copy Jobs, CIFS, and Source Repo

[JoshuaPostSAMC]

Today I have a B&R Server with a primary CIFS repository and I have not set a gateway for this repository so it can pick between the B&R server or any of the 9 proxies I have.
I have a second CIFS repository that I run Copy Jobs to, and on this repository, I have one particular proxy set as the gateway for the repo, so I know all traffic goes through Proxy10.

I'm getting ready to move the second repository and Proxy10 offsite and collecting information to send to the network team for firewalls. When Proxy10 needs to get data from the primary repository, will it only use whatever server is acting as the gateway server, or can it talk only to the B&R server who will get it from the gateway server? Trying to determine if I need the remote proxy to talk to only the B&R server or have the capability of talking to any of the 9 proxies as well. I realize I could change the primary to only use one server as the Gateway server, but I had performance issues in the past like that, so hoping to avoid it.

[foggy]

Connection between gateway servers will be required, so you need to allow all of the source ones.

[JoshuaPostSAMC]

Thanks foggy.

I tried setting my B&R servers as the gateway and my backups went from 200-400 MB/s to 13, so not going to be able to run like that. I'll have to put in rules to allow the remote proxy to talk to all local proxies on the appropriate ports.

In the event that I need to do a restore from the remote repository and proxy, I would still be using the local B&R server to manage that restore. Would the remote proxy need access directly to the local vCenter or ESXi hosts, or would all of that flow through to one of the local proxies? Trying to limit firewall rules, but wanting to make sure we can do an emergency restore without needing to wait for a firewall change request

Any other gotchas I should be concerned with? I assume I would want NFS ports open if I was to do an Instant Restore from the remote proxy. (2 GB link between sites)

[JoshuaPostSAMC]

Second question: If the primary CIFS repository does not have a gateway server defined, when the Remote Proxy10 machine needs to collect data to save to the remote CIFS repository that it will be set at the Gateway Server for, will it for sure select one of the other proxies at the main site to act as the gateway server, or since the primary repository is set to Automatic, could the remote Proxy10 read directly from the CIFS share?

[foggy]

In the event that I need to do a restore from the remote repository and proxy, I would still be using the local B&R server to manage that restore. Would the remote proxy need access directly to the local vCenter or ESXi hosts, or would all of that flow through to one of the local proxies?

In case of full VM restore, Veeam B&R will automatically select the most suitable proxy over which VM data must be transported to the datastore (or you can select the specific proxy to be used). In case of FLR, initial backup mount (with the purpose of browsing files) is performed on the server initiated the restore (backup server or console), while actual restore to original location uses mount either to the backup server or mount server, if it is configured on different server, so make sure the mount server also has appropriate access to the infrastructure.

Second question: If the primary CIFS repository does not have a gateway server defined, when the Remote Proxy10 machine needs to collect data to save to the remote CIFS repository that it will be set at the Gateway Server for, will it for sure select one of the other proxies at the main site to act as the gateway server, or since the primary repository is set to Automatic, could the remote Proxy10 read directly from the CIFS share?

Please review the gateway selection rules in case automatic selection is on.