Removing Snapshots Post-Replication

[xyker]

We're currently using Veeam as one of our use cases to replicate a number of our templates from one site to the other directly to a VMFS datastore. We do not network based storage for our VMs. These templates are basically used to clone and provision an entire software stack for our customers/clients. Due to the large size of one of our VMs (10TB+), whenever we try to clone a VM from one of the replicated templates, the process ends up taking a few hours vs 10 minutes. After some debugging, we've identified that the snapshots created as a restore point, from Veeam on our templates is the cause. After deleting the snapshots completely, the cloning process returned to normal. I've read that VAAI does not work on VMs with snapshots so I think that would explain the slowdown. A few questions:

1) Can we perform replication without any restore points?
2) What are the implications of deleting the snapshots?

[Vitaliy S.]

Hello,

1. No, that's not possible. If you want to just copy templates, then it might be easier to use VM copy jobs instead.
2. Snapshot is required as protection to the corrupted data (malware) which you can replicate from source to the target. Without the snapshot, you wouldn't be able to revert back to the healthy state during a failover operation.

Thank you!

[xyker]

Hi Vitaly,

Correct me if I'm wrong but it appears a VM copy job needs to write my data to some sort of a network file share. Copying a large VM from a network file share over to block storage would take quite a bit of time in our environment.

We're okay with not being able to revert back to a healthy state if we accidentally replicate some bad data. We have other means of restoring our source VMs that we can then re-replicate. Are the snapshots used for any other functionality asides from restoration to a previous state?

[Vitaliy S.]

Correct me if I'm wrong but it appears a VM copy job needs to write my data to some sort of a network file share. Copying a large VM from a network file share over to block storage would take quite a bit of time in our environment.

Yes, you're right. I meant to say file copy jobs as described over here: Replicate Template with Veeam and Job Types. Templates usually don't change frequently, so it seems like a perfect choice for you.

We're okay with not being able to revert back to a healthy state if we accidentally replicate some bad data. We have other means of restoring our source VMs that we can then re-replicate. Are the snapshots used for any other functionality asides from restoration to a previous state?

No, but I guess the replication job would fail and would start calculating digests if you manually delete that snapshot.