Service Provider Architecture Question

[Skyview]

Hello all,

Just getting started with Veeam (v6) and I have an architeture question. I just installed v6 and am currently using it to backup our own internal systems.

However, we just signed up our first customer and I'm wondering whether I use our current v6 backup Server (hosted at our facility) and just use a remote proxy at the customer facility to backup their vmware environment across the wan to a storage location hosted at our facility. Is that the right way to do it? They currently only have about 15 vms, so it's not too big.

Should I be deploying a separate backup server at each remote customer site or only if the size necessitates it?
Is there an advantage to deploying a separate backup server at each remote customer site?
How will reporting/management funtion in both scenarios (1 server at our facility with remote proxies vs remote servers at each customer site)?


Any help would be appreciated. I'll try and answer any additional questions if you have any.

Thanks.

[foggy]

It's sufficient to use the single Veeam backup server/Enterprise Manager to administer such backup infrastructure. There's no need to set up multiple backup servers, a single server can manage all the remote proxies and it can be located in either site. What is more important is to place the backup proxies closer to the data at each customer's site. You are doing this right. Thanks.

[Skyview]

Thanks! One last question-

If I wanted to keep just the most recent version of a vm at the customer site, yet several revisions stored offsite at my facility, is there a way to do that within a single job (point to two different storage nodes)? Or would I need to set up separate jobs pointed to separate storage nodes with separate retention rules?

Thanks, I'll hang up and listen.

[tsightler]

The one "gotcha" I'd point out with using a single server is if you plan to allow the customer to perform self-restores or otherwise have their own access to their backup data. If this is part of your service then you'll need a server per customer.

[foggy]

Each job can have only one target location specified, so you have to use separate jobs in this case or use some external sync tools.

[dellock6]

Thanks! One last question-

If I wanted to keep just the most recent version of a vm at the customer site, yet several revisions stored offsite at my facility, is there a way to do that within a single job (point to two different storage nodes)? Or would I need to set up separate jobs pointed to separate storage nodes with separate retention rules?

Thanks, I'll hang up and listen.

You need to create two separate job, with different rotation schema. Configure the "local" job to be a reverse incremental, so every night you will only inject changed blocks into the previous backup, and keep only one version. Otherwise you will end up in having a full backup every day (bad for the backup window).

Also, remember to separate the execution time of the two jobs, since they will work on the same VMs.

Luca.

[foggy]

No need to care about time actually, as v6 manages overlapping jobs automatically.