Just to be sure, the flow is.

Create VM with root disk on a volume/image.
Create another 1TB volume.
Attach the volume to the VM as a second drive using a multi attach approach (https://docs.openstack.org/cinder/latest/admin/blockstorage-volume-multiattach.html)?
Create snapshot of the original VM (with it's root drive + 1 TB volume)?
Create new VM from snapshot.
Instead of a new VM + 1 new root volume + the same 1TB multiattach volume attached, you get a new multiattach volume (new volume ID and everything)?

I've never had the pleasure to play around with this, but creating a new VM from a snapshot could mean a new volume too. I'm not sure how snapshot will work with multi volume VMs.

On Fri, Sep 24, 2021 at 11:20 AM open infra <openinfradn@gmail.com> wrote:

Hi,

I am planning to spin hundreds of servers with a shared storage.
I have created one instance and attached a 1 TB volume (multiattach) and created a snapshot of the instance.
But when I spin another server using the snapshot, I noticed that instead of expected shared volume, individual 1 TB size additional (dedicated) volume has been attached to each VM.

Is there a mechanism where I can create multiple instances along with a shared volume attached to the instance instead of attaching later?

Regards,
Danishka