[Feature Request] Make CDP usable for Thick Provisioned VMs

[PatrikSE]

The current design of CDP makes it practically unusable for customers using Thick provisioned VMs.

The first part is that during the initial CDP replication, the interim.vmdk inflates to the same size of the base .vmdk if you use disk type Thick Provisioned. After the VM is replicated, it deflates to a more reasonable size. To accommodate the initial interim.vmdk files, the target datastore needs to be twice the size of the source datastore. (There would be workarounds if it were not for the second part of the current design.)

The second part is that the target disk type defaults to Thin Provisioned. There is no way to change the default so if you need Thick Provisioned VMs, you need to use the “pick datastore for selected virtual disks” feature, add each individual VM and change the disk type to Thick Provisioned. The problem is that once you created the job, the entire Datastore part of the job is indefinitely greyed out. If you add any new VMs to the job, you cannot make them Thick Provisioned. You must create new jobs as new VMs are added to the environment. The object from R&D side of blocking the Datastore setting is to prevent change of already replicated VMs, but with the current design it also prevents change of datastore or disk type for any new VMs, not yet replicated.

There are multiple options for Veeam to change the design to make CDP practically usable for Thick Provisioned VMs and I will let up to them to decide which was is the best fix. For me, the best way would be if the interim.vmdk behaves the same regardless of Thin or Thick provisioned VMs and if the default setting for Datastore target was customizable for the job. Also, in the long run I believe a “Veeam Storage vMotion” of CDP replicated VMs would be great so CDP VMs could be moved to different datastores as the environment changes over time.

The above two issues are reported in Veeam Support case #048053 where they were determined as intended design and that I should create a forum post if I wanted it changed in the future.

Hoping for the best.

[wishr]

Hi Partrik,

I apologize for the delayed reply. We'll discuss your request with the teams.

Thanks

[Marcus.A]

Hi,

Sending an update to this FR as it seems it has been fulfilled according to Veeam Helpcenter for V12, https://helpcenter.veeam.com/docs/backu ... ml?ver=120

.. select a type that will be used to restore replica disk files: same as source, thin, thick lazy zeroed or thick eager zeroed.

Edit: It seems this option is only viable for new policies and not existing ones. If I add a new VM to the same policy (which I've selected thick for my first entered VMs) the VM will have the default value of thin provisioned and not thick. It's not possible to edit the existing VMs either so this means I've to create a new, separate policy for the new VM. This process repeats itself for each every new VM I want to add, creating a lot of policies if I don't have them ready when creating the first one.

Is there any changes planned for this behavior or is the workaround to create a new policy for every new VM the accepted solution?

[veremin]

This feature has been since the initial release, but you are right that currently there are certain drawbacks we have not spotted before (ability to set disk type for newly-added VMs, usage of originally configured disk type as a default value for newly-added VMs).

We have tracked this request internally and will address it down the road.

Thanks for raising this!