[XCPBETA] XCP-ng Backup Error After Entire VM Restore/Plugin Feature Requests

[coolsport00]

I created a test XCP-ng Backup Job of a handful of VMs in my test XCP environment. To test Entire VM restore, I deleted a VM from XO. I then restored the VM from my Veeam XCP Plugin Backup Job. All good. But, when my Backup Job ran post VM Restore, the Job couldn't find the restored VM in my Job list. This same behavior occurred in the intial BETA version. My guess is this is because the restored VM gets a different GUID maybe? Can this get resolved?

I went into the Backup Job to re-add the restored VM, then clicked 'Next' to finish the Job wizard. I received a message stating:
"The following objects are missing in the virtual infrastructure. Remove them from Job?", with 2 button options -> Remove or Skip.
If I choose Skip, the VM is retained in the Job VM List and I still get a "failed to find VM" error in the Backup Job, but the same VM I re-added (the restored one) does get backed up..as expected..as a Full. If I choose Remove the Job then no longer displays the error, as I assume this is how Veeam removes the VM from the config DB. I just wanted to share that additional behavior.

I can post logs if needed. Just let me know what specific ones are needed.

Feature requests (minor):
1. Add a Recalculate button in the Backup Job Virtual Machines area to refresh disk sizes
2. Add ability to choose which Worker VMs a Backup Job uses; I know Proxmox doesn't have this feature either...so really wondering at the very least what logic Veeam uses to choose Worker VMs during a Backup for this feature even to be needed?

Thank you.

[Kochkin]

Thank you for the feedback!

VM ID is not preserved on restore to original in beta, that's what cause the original issue. Looks like it is not listed in the forum post yet, I'll ask someone to add it.

Regarding #2, simplified worker selection logic (for Xen backup) is:
- Use any worker from the same node where VM is located (smth like hotadd via NBD)
- If no such worker present, use any worker from the same pool (NBD mode)

"Recommended" configuration is one worker per node (so you would have the best backup speed)

[coolsport00]

Thank you for the info Nikolai. Yeah...would be great for obv reasons if the VM ID is preserved as it is for VMW & HV (not sure if it is in other Plugin products?). So, are you saying it will be preserved if this Plugin goes GA, I assume? Hope so. That would be kind of a big issue

In my testing.. in initial/private BETA and now this Public BETA, I have 2 Hosts with VMs dispersed among them, and 2 Worker VMs. My Backup Job uses only 1 Worker though...even if the VM the Job is backing up is on a different Host than the Worker is on. The Job still seems to use "hotadd" though as seen in the Job task summary. Let me know if you want to see anything.

Thanks.

[Kochkin]

The id should be preserved in ga, that's one of the main things for the restore to original

Regarding worker start logic, we willy air a look. There will be significant change in this logic between beta and ga, that looks like an issue of the beta code. Anyway, worth to take a look from QA for sure

[coolsport00]

Ok, sounds good. Thank you for the additional info Nikolai. Will you be at the Veeam100 Summit in PRG in a few wks?..if so, would like to chat more about the product..just in general.
Best.

[Kochkin]

I actually don’t typically attend the Veeam100, as my role in QA keeps me focused on quality behind the scenes.
However, I’m always happy to discuss the product or answer questions you might have -- feel free to reach out anytime, especially with new beta feedback!

[coolsport00]

Thanks Nikolai

[coolsport00]

Another question, feature enhancement request, if I may Nikolai -> Veeam Replication seems to not be supported; and I think this may be the case also with other current GA Veeam Plugins (Proxmox, KVM, etc)? Will this function be added in a future release/version?
Thanks!