Thanks for the response, that confirmed what I was seeing then. I could only target locally mounted drives on the backup server, or the data domain itself via network proxied through the management server.
Can you elaborate on "adding a gateway server"? Right now the agent only gives me the options to point to the backup server, then pick the repository. There isn't an option to bypass that path for data transfer. I looked through the guide for adding a gateway server, and this states, explicitly, that for dedupe appliances (data domain is what we're using) you have to use fiber channel. I didn't plan on a fiber channel deployment for this location since the core is plenty fast enough, but if fiber is the only way to prevent that traffic from bouncing across the VM hosts, then I'm going to have to add HBAs and connect it.
I'm curious why Veeam is so restrictive in this sense - simply adding a proxy option would seem much less complicated, the same way backups work now with fiber fail-back to network. The ability to target that traffic if it needs to go across the network would be much more flexible than forcing HBAs into the server to accomplish the same task. I couldn't target my mount server to setup the job (I tried), so it makes it a little tough to work around.
Also, I'm curious about one other thing, perhaps you know the answer. If ddboost is intended to move the workload of backup from client to appliance in a direct manner, why does it actually have to pass all that traffic through the mount server anyway? Couldn't it just send metadata to the server and backup directly to the appliance itself as long as there is a route?
Apologies for the questions, but I'm new to the product (but, most definitely not new to backup and recovery) and just working out the kinks in a test deployment and need to have it done quickly. I'm still working my way through the post history for similar questions/issues as what I am seeing.