Problems with restore of large vm

[Spex]

I am currently having problems restoring a vm with several 4 TB disks (16.5 TB in total) using full vm restore. In the restore dialog I specified a separate vmfs datastore for each 4 TB disk (the maximum size of our vmfs datastores is 10 TB - due to the handling and interaction of vms on the same datastore). At this point you even get a warning that the capacity is not sufficient, but you really don't have an explanation and you click next. At the start of the restore, Veeam creates an empty snapshot of the vm and all data ends up in the home directory of the vm. This works fine until, in our case after 30h, the 10 TB are full and we are left behind with an unusable vm that cannot easily be deleted in vcenter.

I opened a case today (Case #05835686) and was told that this is how the design works. There are restore alternatives that are possible depending on the situation and setup, but in my opinion all with limitations:
- Instant recovery - the setup must first be set up (firewall rules, fast kernel interface connection to the storage, storage migration does not handle disks in parallel ...) Our setup is SAN-based so far with the possibility of carrying out fast hot add restores
- Restore the vm files without disks with registration of the vmx file and then restore disk for disk by mapping into new vm. I don't think this corresponds to veeam's claim to be an easy-to-use software. There have also been feature requests on this topic for several years to be able to exclude at least individual disks with the full vm restore - without progress. (veeam-backup-replication-f2/restore-ent ... 42863.html)
- maybe there are some more creative approches out there

You can prepare well for this if you know this limitation - i.e. full vm restore only works for vms that are smaller than the largest data store in total.
But what bothers me is that this is not really communicated anywhere and that with this design (restore into snapshot) the data has to be read and written again after restore step for resolving the snapshot. If you have a fast storage system with xcopy support, this may not be so dramatic, but in other cases a some disaster recoveries can get out of hand quite a bit due to their planed timelimits.

What do you all think about this?

[Origin 2000]

If you use a vSphere Datastore Cluster and a Policy the snapshot behavior changed in the way that they are not created in the VM Folder anymore from a ESXi perspective. Maybe its the same when restore the VM with veeam

Regards,
Joerg

[Spex]

Thanks for this tip. I will ask my vmware team whether this is possible in our setup.

[vmtech123]

Can you not just create a bigger datastore for your restore? 10TB? I can't tell you how often I have to restore 30TB VM's these days.