Wondering why the VBR Appliance has such high resource requirements?

[unsichtbarre]

I am committed to migrating several VBR Windows to VBR Appliance because I believe it is an inherently better (and more secure) path. Nonetheless, I am puzzled by the drastically higher resource requirements for the VBR Appliance. A small/minimal deployment (10 VMs,1 or 2 repos) of VBR on Windows can get away with 2-4 vCPU, 8GB vRAM and 120GB vDISK. The bare minimum for the VBR Appliance is: 8 vCPU, 16GB vRAM and 480GB vDISK, and it seems to want at least 32GB vRAM even on the small deployments. Normally we associate Linux with less overhead/resource use than Windows and this seems to be the opposite?

I also wonder how to scale for a larger VBR Appliance. Say 400-500 VMs (8-14 concurrent jobs of 15-40 VMs each), each job on its own proxy/repo), 12-15 proxies and repos, several SOBRs. Are the minimums for the VBR Appliance adequate for larger deployments or will it require proportionately higher resources?

I am 2 migrations in (out of about 10) total VBR systems. Of course I tackled the smallest/least intense VBR Windows deployments and so far so good. I have some monster VBR Windows to tackle yet, so I guess we'll see. Most of the workload gets transferred to the proxy/data-mover, so experimentation to come.

THX, JB

[scott.anderson]

Probably because all the features and plugins are deployed to the appliance whether you are using them or not. So you will find that all the HyperVisor components are installed for ALL the hypervisors, all the cloud plugins will also be deployed and running. So lots of resource going to functions that you probably wont be using. There are a couple of posts on this affecting the resourcing required to run v13 appliance or Windows instances. There are at least two outstanding requests for Veeam to make the appliance and Windows instances more selectable, but so far Veeam have resisted that functionality.

[eider]

Note that Windows installation recommends 16GB and 8 cores since V13 too, so it's not just VSA that has higher requirements. I'm curious how you arrived at 32GB for the described small environment (10 VMS, 1-2 repos), since that should absolutely work fine with 16GB based on my experience. Make sure to check memory usage in host management Web UI, not in hypervisor since hypervisor will report full memory consumed, along with caches (which differs from how Windows reports it back).

For disk, the minimum requirement is 240GB for OS on VSA and that should be enough for described environment as well. A lot of this size goes into Instant Recovery Cache, so if you do not plan to use it you can size down accordingly, though installer will refuse anything below 240GB.