About backup proxies, transport modes and network switchs

[fcambot]

Hi,

i have 4 physical hosts (vSphere 5.0 essentials plus), 2 vCenters, 2 physical backup servers running VBR 6.5, one virtual backup proxy, and my storage is a stack of dell md32xx enclosures, with SaS connections to hosts (2 Raid cards, each host with one connection to both Raid card).
Being on the process of adding new Nic to vSphere hosts and to "clean up" the switchs, i'd like to "optimize" a bit the whole stuff.
I just watched the "how it works" video, i think i better understand, but i need confirmations .

1) my backup proxy is a virtual machine, and the transport mode is set to "automatic selection". He runs on a host with access to all my datastores, does Veeam uses Virtual Appliance mode in this case ?
2) My "best" backup server has no proxy except himself, should i make a 2nd virtual proxy so i will benefit from Virtual Appliance mode ?
3) If i understood correctly, Veeam communicates with physical hosts using the vSphere management interface, ie the one with the IP@ i can use to login directly with sphere client ?
4) If 3 is yes, so if i plug on the same gigabit switch the network interfaces of : my 2 physical backup servers, my virtual proxy/proxies, the vSphere hosts , the network traffic should be optimal, everybody living on the same floor ?

I hope i've been clear enough.
I truly like VBR, i'm sure i've not yet reached his full potential.

[foggy]

1) my backup proxy is a virtual machine, and the transport mode is set to "automatic selection". He runs on a host with access to all my datastores, does Veeam uses Virtual Appliance mode in this case ?

Yes, it should. You can check that in the job statistics window, just click the particular VM to the left and you should see the [hotadd] tag next to the proxy server's name selected for processing.

2) My "best" backup server has no proxy except himself, should i make a 2nd virtual proxy so i will benefit from Virtual Appliance mode ?

Yes.

3) If i understood correctly, Veeam communicates with physical hosts using the vSphere management interface, ie the one with the IP@ i can use to login directly with sphere client ?

Yes.

4) If 3 is yes, so if i plug on the same gigabit switch the network interfaces of : my 2 physical backup servers, my virtual proxy/proxies, the vSphere hosts , the network traffic should be optimal, everybody living on the same floor ?

What is your backup repository? Local disks on Veeam B&R?
If you are willing to keep backup traffic within one network, this thread might be helpful.

[fcambot]

My repositories are local disks on the veeam servers.
I have [nbd] next to proxy name, does not sound good..
I've manually added datastores, i'll check tomorrow if it changed something.

[foggy]

In fact, manually adding datastores applies to Direct SAN mode only, so this is unlikely to help.

There are several limitations of using hotadd mode outlined in the following KB article: http://www.veeam.com/KB1054, please review.

[fcambot]

I added a proxy for the server which did not have one, and every job worked on hotadd mode .
My 2nd proxy sticks to nbd, i'll dig into this.

I have dedicated a vmnic for each proxy, and regrouped everything involved in backup on same switch.

I think i'm close to the physical server limitations.