understanding Veeam backup proxies

Hi,

Veeam components should only be part of one Veeam deployment for task management reasons.

A backup proxy is used for the reading of data from a hypervisor and then passing that data to a backup repository component to then be written to disk.

If you’ve got a P2P network, are there different subnets or just one stretched L2?

By default your backup job will be set to automatically detect which proxies it can use, this might be suboptimal and your proxy at site a being used for reading from ESXi at site b, you can override this by defining which proxies to use for which jobs 🙂

That sounds correct, as you didn't mention you implemented one separately. During install, Veeam creates a default Proxy (& Repo) role on the Veeam Server itself. It's not ideal because generally, you don't spec the server with enough resources for all the Roles.

What Transport mode do you use? If Hotadd (Virtual Appliance) then you could check the VBR server at Site 2 is actually "doing something" as it would temporarily have the VM disks attached to it during backup processing. 

So each Veeam server is physical? Do you "need" the Site 2 server? No, but what I'd use it for is as a whole separate VBR server to Replicate your Site 1 VMs to Site 2 for DR purposes. Veeam Replication is a petty nifty feature imo. I know you're doing a Copy now, & you could still do so. The Replication is a means by which to recover with really low RTO in the event of some mishap at Site 1. And yes...you'd need to upgrade the Site 2 VBR server to v12 to use it as a component for the Site 1 side. 

Talking again about Proxies...you don't need a physical server for Proxies. You could create a Windows or Linux VM on Site 2, add it as a Managed Server on the Site 1 VBR server, then add it as a Veeam Proxy. 

Just to clarify on this, you don’t use a proxy for a backup copy job, this takes place directly between backup repository components. If you were doing a replication of virtual machines then yes you should use a proxy between each site.

@coolsport00

In normal cases the data mover on the source and target repository will initiate the copy. There are exceptions (repos where a gateway is required).

Wan Accelerator is completely optional.

https://helpcenter.veeam.com/docs/backup/vsphere/backup_copying_process.html?ver=120

https://helpcenter.veeam.com/docs/backup/vsphere/backup_copy_path.html?ver=120

Best,

Fabian

Based on what I’m seeing, you don’t need a full VBR deployment at Site 2, though if you’re using Site2 as DR, I’d suggest moving the VM from Site 1 to Site 2 and run VBR in Site 2 in order to sustain the server still running if Site 1 goes down.  I’d recommend a proxy server at the opposite site of the VBR server.  It’s going to be best to have one at each end to facilitate data flow over the VM network and not use the NBD transport mode over the management NIC’s of the hosts.  I haven’t really tested over a stretched LAN, but I would double check and make sure that when backups run, it’s using the correct local proxy server to facilitate backups.  If not, you’ll need to specify which server to use in the job, but I’m assuming Veeam’s logic isn’t based only on network addresses and actually checks to validate that the backup job will use the local proxy based on accessibility to the local datastores.  Probably not a concern, but I’ve never validated that either.

You wouldn’t run a proxy at the site opposite the VBR server so that it can get direct hotadd access to the vmdk’s on those hosts?  I mean, yeah, you can run it in NBD mode, but that wouldn’t be my preference.