Repository

[Fegume]

Hello.

I am planned installation veeam on environmnet Vmware.

I have 3 nodes esx on cluster
I have one proxy by node
The Store is SAN and the first repository is the same San Store.

I think that better is that the proxy have the repository too, but what node is the better for put in the VM repository Windows?

If I make 3 Repositories (one by node) and use the scale out repository Automatic (the system is the intelligen for use the reposotory correct?)
picture->
https://1drv.ms/u/s!AhEgG4C2IYtVg801UDi ... A?e=X8coaY

[PetrM]

Hello,

1. I believe that proxy affinity rule is exactly what you're looking for.
There is no necessity to create SOBR in this case because data placement policy does not take into account a proxy used by backup job.

2. I don't think it's really important on which node the VM is located as long as you're going to use the single VM as repository.

Thanks!

[Fegume]

PertM thanks for you support!

The proxy affinity will be right if I have more that one repository, buy I only create one repositorie.

If for example I use the proxy02 as repositorie too, the VM backuped with this proxy, will the backup is by FC, but the VM backuped with other Proxy01 or Proxy03 the backup use will the ETH and not FC, don`t it?

[PetrM]

Hello,

The FC will be used for data retrieval as long as proxy works in the direct san transport mode.
The failover to the Network mode (I assume you were meaning this mode by "ETH") will happen by default if proxy fails to use primary transport mode by some reason.
The transport mode selection does not depend on repository location.

The data (already retrieved by proxy) transmission between Source Data Mover and Target Data Mover is always performed through the network channel between proxy and repository.
I'd suggest to take a look at this article to better understand the basic architecture and how different components interact with each other.

Thanks!

[Fegume]

Hi,

Yes ETH (ethernet) = Network

I think that if we use the proxy/repositorie in the same VM, the data mover always send by FC and only use ETH data flow control.

I try explain in this picture:
https://1drv.ms/u/s!AhEgG4C2IYtVg805OWv ... Q?e=Ynkm7j

regards!

[PetrM]

Now, I see. The schema seems to be correct but anyway you may consider deployment with multiple repositories (repository per proxy) and proxy affinity to avoid data flow from Source Data Movers running on proxies 01,02 to the Data Mover running on ProRepo01.

Thanks!