Reorganization of Backup Plan

I don’t see a need for both active and synthetic.  If you’re using a REFS or XFS volume, synthetic fulls should be find due to link block cloning.  If using a different filesystem, you can still use synthetic but know that running synthetic fulls does create a lot of IO and on a NAS, that may not be desired for performance reasons.  Using Active fulls may be preferred because then it’s pretty much just sequential write operations, but know that you’ll be reading the data from the source VM’s so you’ll create more load on your primary storage.  You’ll have to weigh which is better.  Either way, I probably wouldn’t do both.

I’m not going to suggest this as a recommendations, but technically…you can run a docker container (I think) or linux VM (I know) on the NAS itself and present a volume to the VM to be utilized as a Hardened Linux Repository.  I don’t know what performance will be like exactly (I have one client doing this and it works, but they only have a couple of VM’s utilizing this configuration in a copy job), and I’m not sure what a failure would look like exactly, but it can be done.

The better suggestion would be to utilize a physical server (or even workstation) running a Linux distro and connecting to the NAS via ISCSI and utilizing as a hardened repo.  

Finally, one other thing I would look into, but haven’t tried (it was on my list before Veeam released the LHR capability) was to try using the object lock capabilities in MINIO.  This can give you object storage capabilities where you should again be able to utilize immutability.

I’m in the same boat for Linux...I’ve used it here and there over the years, but am by no means any sort of expert.  That said, there’s an ISO I believe now generally available for deploying the LHR’s that make things a lot easier.  I’m about to try it shortly for an internal server.